VE C T O R AN A L YSIS A TE XT-B O O K F O R T H E U SE O F OF M ATH E M ATICS F O UN D E D UP ON JWIL L A R D . P rofe s s o r E D WI N TH E L E C T UR E S OF LL D G IB B S, P H D . f M a th e m a tic a l o B ID AN D P H S TU D E N T S YS ICS . . P by s z c s i n Ya le Un ive rs ity WE L L WIL SO N , PH D . . In s t ru c t o r i n M a t/i n n a t ic s in Ya le Un ive rs ity B N EW YO RK : C H ARLE S S CRI N ER ’ S S O N S LO N DO N : EDWARD ARN O LD 1 901 B Y YA L E P U u bl i s h e d , U NIV E S I R De c e mb e r , 1 9 01 TY . . NIV ITY PCAM B IDJOG N IL ON AN D ERS SO N H RESS R W E, S U S A . . . P RE FA CE B Y P R OFE SSOR G IB B S the printing of a short pamphl et on the E le m e n ts of Ve c t o r An a ly s is in the years 1 88 1 -8 4 never published but somewhat widely circulated among those who were known t o the desire has been e xpressed b e interested in the sub ject in more than one quarter that the sub stance of that t rea tise perhaps in fuller form should be made accessib le t o the pub lic A s however the ears passed without my fi ndi ng the leisure to meet this want which seemed a real one I was very glad to have one of the hearers of my course on Vec t or A nalysis in the year 1 8 9 9—1 900 undertake the preparation of a tex t-b ook on the sub ject I have not desired t hat Dr Wilson should aim simply at the reproduction of my lectures but rather that he shoul d use his own judgment in all respects for the production of a te x t b ook in which the subject should be so illustrated b y an adequat e numb er of e x amples as to meet the wants of s t u dents o i geometr and ph sics SIN C E , , , , , , y . , , , , . . , - y NI ER S ITY YALE U V y , Se p t e m . JWILLARD . be r , 1 9 01 . ENG INEERIHG l BB GI S . G EN E RA L P REFA CE WH EN I undertook to adapt the lectures of P rofessor G ibbs on VE CTOR A N A LY S IS for pub lication in the Yale B ic e n t e n nial Series Professor G ibb s himself was already so fully engaged upon his work to appear in the same series E le m e n t a ry P r in c ip les i n St a ti s ti c a l M e c h a n ic s that it w a s understood no material assistance in the composition of this book could b e e xpected from him Fo r thi s reason he wished me to feel entirely free to use my own di scretion ali ke in t h e selection of the topics to b e treated and in t h e mode of treatment It has been my endeavor to use the freedom thus granted only in so far as was necessar for presenting his m e t hod in te x t-book form B y far the greater part of the material used in th e follow ing pages has been taken from the course of lectures on V ector A nalysis delivered annually at the Universit by Professor G ib bs Some us e however has been made of the chapters on V ector A nalysis in Mr O liver H e a vis ide s E le c t ro m a gn e tic Th e o ry ( Electrician Series 1 8 9 3 ) and in P rofessor Fop p l s lectures on D ie M a xwe ll s c h e Th e o ri e de r E le c t ric i t ci t T n s has u b ner e M previous study of uaternio Q ( y also b een of great assistance Th e m aterial thus o b t ained has been arranged in the way whi ch seems b est suited to easy mastery of the sub jec t Those Arts whi ch it seemed b est to incorporate in the te x t b u t which for various reasons may well b e omitted at the fi rst readi ng have b een marked with an asterisk Nu m e ro u s illustra t ive e x amples have been dra wn from geometry mechani c s and physics Indeed a large part of the te x t has to do wi t h applications of the met hod These application s have not been set apart in chapters b y themselves b u t have , , , . . y . y . , , ’ . , ’ ’ ’ , . . . , , . , . , G E N E RAL P RE FA CE b een distributed throughout t h e b ody of the book as fast as the analysis has b een developed su ffi ciently for t heir adequate treat men t It is hoped t hat b y t his means t h e reader may b e G rea t care b e t t er enab led t o mak e prac t ical use of the book has b een t a k en in avoidi ng t h e introduction of unnecessary ideas and in so illus t ra t ing each idea that is introduced as to mak e its necess it y evident and its meani ng easy to grasp Thus t h e b ook is not intended as a complete e x position of the t heory of Vector Analysis b u t as a te xt-b o ok from whi ch so much of t h e sub jec t as may b e required for practical a p p li cations may b e learned Hence a summar inclu di ng a list of t h e more important formul ae and a num b er of e x erc ises have b een placed a t t h e end of each chapter and many less e s sen t ial points in the te x t have b een indica t ed rather t han fully worked o u t in t h e hope that t h e reader will supply th e de t ails Th e sum mary may b e found useful in revie ws and for reference Th e s ub ject of Vector A nalysis naturally divides i t self into three di stinc t part s Firs t t hat whi ch concerns addi t ion and the scalar and vector produc t s of vectors Second that which concerns the differential and i ntegral calculus in it s relations to scalar and vector func t ions Third that which contain s the theory of the lin ear vector function Th e fir st part is a necessary introdu c t ion to b oth other parts Th e second and t hird are mutually independent Ei t her may be taken up fi rst Fo r practic al purposes in mathema t ical physics the second must b e regarded as more elementary than the t hird B u t a studen t not prima ri ly in t ereste d in physics would nat u ra ll fi pass from the rst part t o the t hird which he would y pro b ab ly fi n d more attrac t ive and easy than the second Follo wi ng thi s division of t h e sub ject the main b ody of the b ook is divided into s ix chap t ers of which two deal wi t h each of the three part s in the order named Chapters I and II treat of addi t ion su b trac t ion scalar mul t iplica t ion and t h e scalar and vector products of vectors Th e e x po s i t ion h a s b een made quite elementary It can rea di ly b e under stood b y and is especially suited for such readers as have a k nowledge of only the elements of Trigonometry and An a . . , . , y . , , , , , . . . , . . , , . . . . . , . , . . , , , . . . y lytic G eometr Those who are w ell versed in Q uaternions or allied sub jects may perhaps need to read only t h e s u m m a ries Cha pters III and IV c ontain the treat ment of those topics in Vecto r A nalysis which t hough of les s value to the studen t s of pure mathematics are of the utmost impor t a nce t o students of physics Chapters V and VI deal wi t h the linear vector function To s t udents of physics t h e linear vector function is of particular importance in the m a t h e m a t i cal trea t ment o f phenomena conn ected with non isotropic media ; and to the student of pure ma t hemati cs t his part of the b ook will prob ab ly b e the most interesting of all o win g t o the fact tha t it lea ds to M ulti ple A lge b ra or the Theory of Matrices A concluding chapter VII which con t ains the development of cert ain hi gher parts of the theory a num b er of applica t ions and a short sketch of imaginary or comple x vectors has b een added In the treatment of the integral calculus Chapter IV questions of mathematical rigor arise Al t hough modern theoris ts are devoting much time and thought to rigor and al t hough they will doub tless cri t icise this portion of the b ook adversely it has been deemed b est to give b u t li ttle attention to t h e disc u ssion of this sub ject An d the more so for the reason that whatever system of notation b e employed ques tions of rigor are indissolub ly associate d wi t h t h e c a lc u lus and occasion no n e w diffi cul t to t h e s t udent of V ector A nalysis who must fi rst learn what the facts are and may pos t pone until later the detailed consideration of the re s t ric tions that are put upon those facts N otwi t hstan di ng t h e e fforts which have b een made dur ing more than half a century to introduce Q uaternions in to physics the fact remains that t hey ha ve not fou nd wi de favor O n the o t her hand there has b e en a growing tendency espe c ia ll y in the l a st decade toward the adoption of some form of V ector A nalysis Th e works of Heaviside and Fo p p l re ferred to b efore may be cited in evidence A s yet ho wever no system of V ector Analys is whi ch makes any claim to completeness has b een publ is hed In fact Heavi side s ays : I am in hopes that the chapter whi ch I now fi ni sh may . . . . , , . . . . - , . . , , , , , . . , , . , , . y , . . . . . “ G E N E R AL xii P R E FA C E serve as a stopgap till r e gular vectorial treatises come to be wri t t en suitab le for physicis t s b ased upon the vectorial treat ment of vectors ( Elec t romagne t ic Theory Vo l I p E ls e where in the same chapter Heaviside has set fort h the claims of vector analysis a s against Q uaternions and others have e xpre s s ed similar views Th e k eynote then to any system of vector analysis must This I feel con fi dent was P rofessor b e its prac t ical utility He us es it G ib bs s po int of view in b uil di ng up his system entirely in his cours es on Electricit y and M agne t is m and on ight In writing this b ook I Elec t romagne t ic Theory of have t ried to present the sub ject from this prac t ical s t and point and keep clearly before t h e reader s mind the ques t ions : What comb inations or functions of vectors occur in physics and geometry ? A n d how m a y these be represented symb olically in t h e way b es t sui t ed to facile anal t ic manip Th e t reatmen t of these que s t ions in modern b ooks ula t io n ? on phys ics has b een too much confi ned to the addi tion and sub t raction of vectors This is scarcely enough It has b een the aim here to give also an e x position of scalar and vector products of the Operator v of divergence and curl which have gained such universal recognition since the a p p e a ra n c e of M a x well s Tre a t is e o n E le c tri c i ty a n d M a gn e tis m of slope po t ential linear vector function etc such as shall be adequat e for t h e needs of students of physics at the present day and adapted to them It has b een asserted by some that Q uaternions V ector A nalysis and all such alge b ras are of little value for investi ga t ing ques t ions l n mathematical physics Whe t her this assert ion shall prove true or not one may s t ill maintain that vectors are to ma t hematical phys ics what invariants are to geometry A s every geometer must b e thoroughl y conver sant with the ideas of invariants so every student of ph sics shou ld be ab le to t hin k in terms of vecto rs A n d there is no way in which he espe c ially a t t h e b eginning of his sci e n t ifi c studies can come to so true an appreciation of t h e import an c e of vecto rs and of the ideas connected with t hem as by working in Vector A nalysis and dealing directly with , . , . . , , . , , , , . ’ . L . ’ , y . . , , ’ , , , . , , . , , . , y . , . , , GE N E RA L r P RE FA CE i x ii the vecto s themselves To those that hold these views the success of P rofessor Fop p l s Vo rle s u n ge n itbe r Te c h n is c h e M e c h a n i le ( four volum es Te u b n e r 1 8 9 7 —1 9 00 already in a second edi t ion ) in which the t heory of mechanics is devel oped b y means o f a vector analysis can be b u t an e n c o u r aging sign I t ake pleasure in tha nking my colleagues D r M B P o rt e r and P rof H A B umstead for assist in g m e wit h t h e m a nu script Th e good servi ces o f the latte r have been particularly valuab le in arrangin g Chapters III and IV in their present form and in suggesting many of the illustrations used in the work I am also un der ob ligations to my father Mr E dwin H Wilson for help in connec t ion both wi t h the proofs and the manuscript Finally I wish to e xpress my deep indeb t Fo r although he h a s been so e dn e s s to P rofessor G i bb s preoccupied as to be u nable t o read ei t her manuscript or proof he h a s always b een ready to talk matte rs over with me and it is he who has furnished me with inspiration suf fi c ie n t to carr through the work . ’ , , , , , . . , . . ' . , . . . . . . , . , . , . . , , y . ED YA L E U NI E S I V R TY , O c to b e r , 1 901 . WIN B ID WE LL WIL S O N . TA B LE O F C O N TE N TS REFAC E BY ROFE SSOR GIBBS G ENERA L REFAC E P P P CHA P TE R I AD D ITION A ND S M ULTIP LICATION S C ALAR CAUALALR SANDANDNULLECTOR S VECTOR S P OINT OF VIEW OF T I S C AP TER CADDITION L AR MU L TI PL ICATION NE G ATIV E S IGN P A A LL E L OGRAM L AW UAWB TRACTION S G OVERNING T E FOREGOING O P ERATION S OM PONENT S OF VECTOR S ECTOR E UATION S T RE E UNIT VECTOR S PPL IC A TION S S UNDRY P RO B L EM S IN G EOMETRY E CTOR RE L ATION S INDE P ENDENT OF T E ORIGIN ENTERU S ES OFOF GRAVITY ARYC E NTRIC CO RDIN A T E S V E CTOR S TO DENOT E AREA S U MARY OFON CC AAPPTERT E R I I V EQ TH S H E S TH . A TH . E H E R H L C . TH E H V i, A Q jk , To . H V C B . TH O E S H H M E xE R CI SE S CH AP TE R II D IRE C T 2 7 28 TH E 2 9—3 0 TH E 3 1 —3 3 TH E 3 4—3 5 TH E 36 TH E AN D S KE W P R OD U C TS OF VECTORS DIRECT S CA LA R OR DOT P RODUCT OF DIS SEWTRIBVECTOR UTIVE L AWOR CRO A ND APPL ICATION S SS PRODUCT OF DITRIS TRILEB UTIV E L AW A ND A PPL ICATION S P RODUCT B , , K , P , A ° C Two Two VE CTOR S VECTOR S 55 58 60 63 67 C ON TE N TS x vi CA AR TRI PLE P RODUCT B OR VECTOR TRI PL E P RODUCT B RODUCT S OF MORE T AN T REE VECTOR S WIT CATION S RIGONOMETRY ECI P ROCA L S Y S TEM S OF T RE E V E CTOR S OLAUTION OF S CA L AR AND V E CTOR E UATION S L INE A R IN N UN NOWN VECTOR YINEMATIC S TEM S OF FORCE S A CTING ON A RIGID B OD S OF A RI G ID B ODY ONDITION S FOR E UI L I B RIUM OF A RI G ID B ODY ETLATION S B ETWEEN OF J REE P ER P ENDICU L AR UNIT VECTOR S ROUMMARY B L EM S IN G EOMETRY L ANAR CO RDI ATE S OFONC C AAPTER II P ER II S L TH E TH E A x0 [ A B C] A X ( X C) ° H P To H A P P LI T H R S H Q K S Y K C Q T wo R R I G H TZH Y ST E M S AN DE D H P S H H E xE R C ISE S O P . N T C H AP TER III TH E 5 5— 5 6 DIFFE REN TIAL C ALCUL U S OF VE C TOR S RIVATIVE S AND DIFFER E NTI AL S OF VECTOR FUNCTION S A S CA L AR ARIA B L E URVATURE A ND TOR S ION OF GAUC E CURV S INEMATIC S OF A P ARTIC LE ODOGRA P IN S TANTANEOU S OF ROTATION NTE G RATION W T APPL IC A TION S INEM A TIC S CALARECTOR FUNCTION S OF PO S ITION IN S PACE DIFFERENTIATING O P ERATOR S CA L A O P ERATOR ECTOR FUNCTIONS OF POSITIO PA E NTER P RETATION OF T E NTER PR E TATION OF T E CUR L AWSPOFARTIAO PERATION OF L A PPL ICATION OF OF A VEC TOR FUNCTION A NALOGOU S AYLOR S EOREM DIFFERENTIATING O P ERATOR S OF T E S ECOND OR E R G EOMETRIC INTER PRETATION OF A PLA C E S O P ERATOR T E DI S PE R S ION UMMARY OFONC C AAPTER III PT E R III DE W C E SP EQ . V To K . TH H TH E H E H A x rs E I I H To K S TH E TH E V A jz R o V N H H I I L E 78 TH V o . T ’ H E L V V Q Vx VX V V V E x P A N SIO N To 77 § D IV E R G ENQ E , TH IN As S Ex E R CISE S H H H TH . D ’ 1 59 1 66 C ON TE N TS XVII CH AP TE R IV A INTIONE INTEGRA L S OF VECTOR FUNCTION S WIT APPL ICA S G AU SS S EOREM ONVER S E OF TO E S S EOR E M WIT APPL ICATION S RAN S FORMATION S OF LINE S URFACE AND VOL UM E EMAR S ON M U L TI PL E VA L UED FUNCTION S OTENTIA L INTEGRATING O P ERATOR OMMUTATIVE PRO P ERTY OF AND EMARINTES GURATIN P ON T E FOREGOING G O P ERATOR S E LATION S B OETWEEN T E INTEGRATING AND DIFF ER P ERATOR S P OTENTIA L A S O L UTION OF OI SS ON S UATION OFUNCTION L ENOIDA L AND IRROT A TIONA L P ART S OF A VECTOR ERTAIN O P ERATOR S AND T EIR INVER S E M UTUA L OTENTIA L S EWTONIAN S APL ACIAN S AND ERTAIN B OUND A RY VA L UE T EOR E M S UMMARY OF C A P TER IV E RCI S E S ON C A P TER IV TH E INTEGRAL C ALC UL U S OF VE CTORS P H L TH ’ C S TH ’ K T H , R - K P TH . “ E C H K TH POT V P OT R IN , “ E N E W, ” “ LA P , ” “ MAx H R E N TIA T I N G TH “ E POT ” ’ P Is EQ 9 3 94 S ‘ . C H P N , , L , M A X WE L L IA N S C S H H H EX CHA P TE R V LINE AR VE C TO R F UN C TIONS INE A R VECTOR FUNCTION S DEFIN E D D YADIC S DEFINED L IN E AR VECTOR FUNCTION MAY B E RE PR E S ENTED B Y A DYADIC RO P ERTIE S OF DY A DIC S NONION FORM OF A DYADIC DYAD OR INDETERMINATE P RODUCT OF VEC TOR S I S T E MO S T GENERA L UNCTIONA L P RO P ERTY OF T E S CA L AR AND VECTOR P RODUCT S RODUCT S OF DYADIC S D EGREE S OF NU LL IT OF DYADIC S IDEMF ACTOR L AN Y . TH E TH E P Two H . H P Y TH E F GE II C ON TE N TS 0 0 . XVI ECI P ROCA L DY A DIC S OWER S A ND ROOT S OF DYADIC S ONS ELUFGCON ATE U GDYADIC S E L F CON U G ATE AND ANTI ATE P ART S OF A DYADIC NTI S ELF UADRANTA CON UGATEL DYADIC S VECTOR PROD EDUCTION OF DYADIC S NORMA L FORM D OU B L E MU L TI PL ICATION OF DYADIC S S ECOND AND T IRD OF A DYADIC ONDITION S FOR DIFFERENT DEGREE S OF NU LL ITY ONION FORM D ETERMINANT S NVARIUATION A NT S OF A DYADIC H A MI L TON A Y LE Y UMMARY OF C A P TER V ERCI S E S ON C A PTER R C J A - UCT . S . J - J - 1 1 2 -1 1 4 P . J - TH . Q E V E R SO R S R To TH H E C N . I TH . C - E EQ S H H EX v CH AP TE R VI ROTATION S AN D S TRAIN S OMOGENEOU S S TRAIN RE PRE S ENTED B Y A DY A DIC OTATION S A OUT A FI ED POINT VECTOR S EMI TANGENT OF VER S ION BIYCUADRANTA L AND T EIR P RODUCT S L IC DYADIC S IGONICTS TENANDSOR S EDUCTION OF DYADIC S TO CANONICA L FORM S TONIC S S IM PL E AND S EARER S UMMARY OF C A PTER VI H R X B TH . VE R SO R S - E H V E R SO R S Q C H R T C Y C L O TO N I C S R , C o M P L Ex C Y C L O TO N I C S , S H H CH AP TE R VII MIS C E LL AN EOU S AP P LICATION S 1 36 1 42 1 4 3 -1 4 6 1 4 7— 1 4 8 1 4 9 —1 5 7 1 5 8 -1 6 2 UADRIC S URFACE S P RO PAG ATION OF IG T IN CRY S TA L S ARIA B L E DYADIC S URVATUR E OF S URFACE S H A MONIC I B RATION S AND Q TH L E H V C R V B IV E C T O R S , c i Cf / t VE C TOR A N AL YSI S , Th e p o s i ti ve typ ic a l s c a l a rs a n d n e ga tive n u m be rs o f lge bra a re the alge b ra is called o rdi n a ry a this reason th e ordinary t o distinguish it from the s c a la r alge b ra when necessar algeb ra or analys is which is the sub ject of thi s book Fo r . y ve c to r . Th e typ ic a l ve c to r is the disp la C onsider fi rst a point P ( Fig f c emen t o L et . t ra n s la tio n i n sp a c e . P b e di splaced in a straight line and take a new position P This change of position is represented by the Th e magnitude of the di splace 1 line P P ment is the lengt h of P P ; the direction of it is the direction of the li ne P P fro m P to N e x t cons ider a displacement not of one P F G 1 but of all the points in space et all the points move in s t raight li nes in t h e same direction and for the same dista nce D This is equivalent to shifting space as a rigid b ody in t hat direction through t h e dis t ance D without ro t ation Such a displacement is called a tra n s la tio n It possesses di rection and magnitude When Space undergoes a translation T each point of Space undergoes a displacemen t equal to T in magnitude and d rection ; and conversely if t h e displacement P P which any one particular point P s u f fers in t h e translati on T is known then that of any other point Q is a so known : for Q Q must b e equal and parall el to P P Th e translation T is represented geometrically or graphically by an arrow T ( Fig 1 ) of which the magn itu de and direction are equal to those of the translation Th e ab solute position of th is arrow in Spa ce is entirely immate rial Technically the arrow is called a s tro ke Its tail or initial point is its o rigi n ; and its he a d or fi nal point its te rm in u s In the fi gure the origin is designated by 0 and the terminus b y T Thi s geo metri c quanti t y a stroke is used as the mathematical sym b ol for all vectors jus t as the ordinary positive and negative num bers are used as the symbols for all scalars ’ . ' ' ~ . ’ ’ ’ L . I . . . , . . . . i , ’ l , ’ ’ . . . . . , . . , , , . A DDI TION AND 3 S CA LAR M UL TIP LI CA TI ON ex amples of scalar quantities mass ti me den O thers are di s s ity and temperature have been mentioned tance volume moment of inertia work etc Magnitude however is b y no means the sole propert y of these quantities E ach has its Each implies somethi ng b esides m a gni t ude own d stingu ishing characte ristics as an e xample of which its dim e n s i o n s in the sense well known to physicists may be cited A distance 3 a time 3 a work 3 etc are ver differen t Th e magnitude 3 is ho wever a prope rty common to them all — perhaps the only one O f a l l scalar quanti tities p u re n u m be r is the Simplest It implies nothing but magnitude It is the scalar p a r e xc e lle n c e and consequentl it is used as the mathemat ical symbol for all sca ars A s e x amples of vector quantiti e s force displacement velo city and acceleration have been given Each of these has other characteristi cs than those which b elong to a vecto r pure and Simple Th e concept of vec to r involves t wo ideas and two alone — magnitude of the vector and direction of the vector B u t force s more complicate d When it is applied to a rigid bod the lin e in which it a cts must b e taken into consideration ; magnitude and di rection alone do not suf An d in cas e it is applied to a non -rigid bo dy the p o i n t fice of application of the force is as important as the magnitude or Such is frequen t ly true for vector quantit ies other direc t ion tha n force M oreover t h e question of dimensions is present as in the case of scalar quantities Th e mathematical vector t h e s t roke which is the primary object of consideration in t hi s b ook a b s t racts from all directed quantities their m agni tude and di rection and nothing but t hese ; jus t as the mat he m a ti c a l scalar pure numb er abstracts the magni t ude and that alone Hence one must b e on his guard lest from analogy he attribute some properties to the mathematical vector which do not belong to it ; and he must be even more careful lest he Obta in erroneous results b considering the As , , . , , , , , . , , . i . , , . , , , . y , , . y . l . , . , . , . y . i . . . . . , , , , , . y 4 VE C TOR A N AL YSIS vector quantities of physics as possessing no properties other than th o se of the mat hemat ical vector Fo r e x ample it would never do to consider force and its e ffects as unaltered b y s h ift ing it parallel t o it s elf This warnin g may not be necessary yet it may pos s ib ly s ave some confusion Inasmuch as taken in its entirety a vector or stroke is b u t a Single concept it may appropriately be designated b y one let ter O wing however to the fundamental di fference be tween scalars and vecto rs it is necessary to dis t inguish caref u lly t h e one from the o t her Sometimes as in mathe m a t ic a l physics t h e distinction is furnished b y the physica int erpreta t ion Thus if n b e the inde x of refraction it must b e scalar ; m the mass and t the time are also scalars ; b u t f t h e force and a the acceleration are vectors When however t h e letters are regarded merely symb ols with no particular physical signi fi cance some as t pographical difference must b e relied upon to distinguish vectors from scalars Hence in this b ook Clar e n do n typ e is used for sett ing up vectors and ordinary type for scalars This permits t h e use of t h e same letter di fle re n t ly printed to represent the vector and its scalar magnitude } Thus if G b e t h e electric current in magni t ude and direction 0 may b e used to represent the magnitude of tha t current ; if g b e the vector acceleration due to gravity g may be the s calar value of that acceleration ; if v b e the veloci t y of a moving mass v may b e the magni t ude of that velocit y Th e use of Clarendons to denote vecto rs makes it possi b le to pass from directed quan t i t ies to t heir scalar magnitudes by a mere change in t h e appearance of a letter wi t hout any confusing change in the lette r itself De n i t io n Two vectors are said to b e e qu a l when they have the same magnitude and t h e same direction . . . , , , , . , l , . , . , , , , , , , . , , , y . . , , , . fi i m . by n o e a n s i n va ria bl y fo llo we d In in s t a n c e s it wo ul d p ro ve j us t a s u n d e s i ra bl e a s it is c o n ve n i e n t in o t h e rs c h i e fly va l ua bl e in t h e a ppl i c a t i o n o f ve c t o rs t o p h y s i c s 1 Th i s . c o n ve n t o n , h o we ve r, is . . . so m e It is A DDI TI ON A N D y of two vectors equali t usual Sign Th e Thus 5 S CA LAR M UL TIP LI CA TI ON A and A B B is denoted by the . y Evidently a vector or stroke is not a ltered b shi fting it ab out p a r a lle l to itself in space Hence any vector A P P ( Fig 1 ) may b e dr a wn from any a s signed point 0 as origin for t h e segment P P may b e moved parallel to itself unt il the point P fal ls upon the poin t 0 and P upon some po nt T ’ . . ’ i ’ . Th e n y this wa all vectors in space may b e replaced by directed segments radiating from one fi x ed point 0. E qual vectors in space will of course coincide when placed with their ter mini a t the same point 0 Thus ( Fig 1 ) A P P and B Q Q b o th fall upon T 0T Fo r the numerical determination of a vector th re e scalars are necessary These may be chosen in a variet of ways 0 b e polar co ordinates in Space any vector r drawn If r wi t h its origin at th e origin of co ordinates may be represented by the three scalars r (I) 9 which determine the terminus of the vector 7 S In , ’ ’ . , . , . y . . , ’ , , ( . q, , if x y z be Cartesian coordi na te s in space a vector r may be considered as given b y the differences of the co ordinates as y z of its term inus and those at y z of its origin Or , , ’ , ’ ’ , r - co , y , , r - . y , z r i in particul ar the origin of the vector co ncide with the origin of c o Ordi n a t e s the vector will b e represented by the three co ordin ates of its termi nus If , l (m r , y" 2 1 ) r When tw o vecto s are equal the three scala s which re p e sent them must b e equal respecti vely each to each Hence one vector equalit implies th ree scalar equali t ies y r . . 6 fi VE C TOR A N AL YS I S q l vector A is said to be e ua to z e ro when its magnit ude A is z ero Such a vector A is called a n u ll or z e ro vec to r and is writte n equal to naught in the usual manner Thus De A n i ti on . . A = O if A = 0 . i null vectors are regarded as equ al to each other w thout any considerat ions of direction In fact a n u ll vec t or from a geometrical standpoint woul d that is to b e represented b y a linear segment of length z ero say b y a point It consequently woul d have a W holly inde terminate direction or what amounts to t h e same thing none at all If however it b e regarded as the limit approached b a vector of fi n i te length it might be considered to have that d rection which is the limit approached by the direction of the fi ni t e vector when t h e lengt h decreases inde fi ni t ely and a p T ero as a limit h ju t i fi cation for sregar ng r a h z e s di d i s o c e p this direction and looking upon all n ul l vectors as equal is tha t when they are added ( Art 8 ) to other vectors no change occurs and when mul t iplied ( Arts 2 7 3 1 ) b y o t her vectors the product is z ero In e x tending to vectors the fundamental Opera t ions of algebra and arithmetic namely addition sub traction and multiplication care must be e x ercised not only to avoid self contradictor de fi nitions b u t also to lay down u seful ones B oth these ends may be accompl ished most naturally and easil y b y looking to physics ( for in that science vectors con t in u a ll y present themselves ) and b y observing how such quantiti es are treated there If then A be a given displace ment force or velocity what is t wo three or in general 3: t imes A ? Wha t the negative of A ? An d if B be ano t her what is the sum of A and B ? That is to say what is the equivalent of A and B taken together ? Th e obvious answers to these questi ons suggest immedi ately the desired de fi nitions All . , . . y , , , , i , , . . . , . y , , , , , . . , , , , , , , , . A DD I TI ON fi A ND Sc a la 7 S CA LAR M UL TIP LI CA TI ON r M u ltip lic a t io n y vector is said to be mul tiplied b a ositive scalar when its magni t ude is mul tiplied by that scalar and its direction is left unaltere times Thus if V be a velocity of nine knots E ast b y N orth v is a velocity of t went -one k nots with the di rection still O r if f be the force e x erted upon the scale E ast b y N orth pan b y a gram weight 1 000 times f is the force e x erted b y a kilogram Th e direction in both cases is verti cally down ward If A be the vector and a: the scalar the p oduct of ac and A is denoted as usual by a: A or A cc De p n i tio n A d . y , . , . r . . y is however more customar to place the scalar multiplier before the multipli cand A This multiplicati on by a scalar is called s c a la r m u ltip lic a tio n and it follows the assoc iati ve law It , , . , y as in or di nar algebra and arithmetic This statement is im me di atel ob vious when the fact is taken into consideration that scal a r mul tiplication does not alter directi on but merely multiplies the length De n i tio n : A u n i t vector is one whose magnitude is uni t A n y vector A may be looked upon as the product of a unit vector a in its di rection b y the positive sca ar A its magni tude y . fi y . l . , . A A a unit vector a may similarly be written as the product of A by l /A or as the uotient of A and A Th e q . 8 fi VE C TOR A N A L YS I S negative Sign pre fix ed to a vector re ve rs e s its di rection b u t leaves its magni tude un changed Fo r e x ample if A b e a displacement for two feet to t h e right A gain if the A is a di s placement for two feet to the left s t roke A B b e A the stroke B A whi ch is of the same length as A B b u t whi ch is in the direction from B to A instead of from A to B will be A A nother ill ustration of the use of t h e negative S gn ma y b e taken from N ewton s th ird la w of motion If A denote an action A will denote t h e reaction may be prefi x ed to a vec Th e positive Sign tor to call particular at tention to the fact that t h e direction has not been reversed Th e two signs and when used in connection wi t h scalar m ul tiplication of vectors foll ow the same laws of operation as in ordi nary algebra Thes e are symboli cally De n i tio n : Th e , . , . , , , i . ’ ” “ , . ” . , . . — Th e interpretation is ob vious A ddi tio n A) . . a nd Su btra c ti o n addi tion of t wo vectors or strokes may be treated mos t simply b y regarding them as de fi n ing translatio n s in space (Art et S b e one vector and T the other et P be a poin t of space ( Fig Th e trans lat ion S carries P into P such that the li ne P P is equal to S in magni tude and direction Th e transformation T will then ” carr P into P the line P P b eing parallel to T and equal to it in magnitude Consequently the result of S follo wed by Fm 2 T is to carry t h e point P into the point P If now Q be any other point in space S will carry Q in t o Q such that QQ S and T will then carry Q into Q Th e L . L . . ’ ’ y . ’ ’ ” . . , ” . , ’ ’ ’ ” 10 s VE C TOR A N A L YS IS q y p e uentl the oints r and P lie at the ve tices of a parallelogram Hence ” is equal and par P P allel to P P Hence S T fol carries P into P lowed b y S therefore car ries P into P through P whereas 8 followed b y T carries P into P thr ough Th e fi nal result is in This ma be designated s mbolicall P, P ’ , P ’” ” , . ” ’ ’ . ” ’” ‘ . ’ ” , ” FI G 3 . . either case the same b y writing . y y y is to be noticed that S P P and T P P are the two sides of the parall elogram P P P P which have the point P as common origin ; and that R P P is the diagonal dr awn through P This leads to another ver common wa of stating the defi nition of the sum of two vectors If two vecto s be drawn from the same origin and a parall elo gram be constructed upon them as sides their sum will be that diagonal which passes through their com mon origin parallelogram law according to Thi s is the well-known whi ch the physical vector quantities force acceleration v o loe ity and angular velocity are compounded It is important to note that in case the vectors lie along the same line vector addi tion b ecomes equivalent to algebraic scalar addi tion Th e lengths of the t wo vectors to b e added are added if the vectors have the same direction ; but subtracted if the have Oppo site directions In either case the sum has the same d rection as that of the greater vector After the de fi n ition of the sum of two vecto s has b een laid down the sum of several ma be found b y adding t o gether the fi rst two to this sum the third to this the fourth and so on until all the vectors have been combined into a sin ’ ’ It ” ’ ’” ” y . y r . , . ” “ , , . , . y . r . y , , i , , A DDI TION AND ALAR SC M P UL TI LI CA TI ON 11 gle one Th e fi nal result is the same as tha t obtained by placing th e origin of each succeedin g vector upon the terminus of the precedin g one a n d then drawing at once the vecto r from the origin of the fi rst to the term inus of the last In cas e these two points coincide the vecto s form a closed polygon and t heir sum is z ero Interpreted geometrically this s t ates that if a number of dis placements R S T are such that the strokes R S T form the sides of a closed polygon taken in order then the effect of carr ing out the displacements is n il E ach point of space is brought b ack to its starting point In t e rp re t e d in mecha nics it state s that if an number of forces form the sides of a closed polygon a c t at a point and if the taken in order then the resulta nt force is z ero and t h e po int s in equi librium under the action of the forces Th e order of sequence of the vectors in a sum is of no con sequence Thi s may b e shown by proving that any two a dja fect ng the resu t c e n t vecto s may be interchanged w thout af To show . r . . , , , y , , . y y i . , . . r L et hen L et now conse quently i i T Then 6 B 0 1) B O = D ” ” . 0 D =C ’ . l . p r ll l r is a a a e og am and Hence i y wh ch proves the statement Since an two adjacent vectors may be interchanged and since the sum may be a anged in an y order by successive interchanges of adjacent vectors the order in whi ch the vectors occur in the sum is immaterial De n i t io n : A vector is said to be subtrac ted when it s added afte r reve sa of di rection S mbolically . rr , , fi i By i . rl A r i . B A th s means subt act on is reduce y , B) . d to ad di tion and needs VECTOR ANA L YSIS 12 no special consideration There is however an interestin g and import an t way of represen t ing t h e di fference of two vec t ors C omplete geome trically et A 0A B O B ( Fig the parallelogram of whi ch A and B are t h e sides Then the di agonal O O = C is the sum A + B of t h e two vectors N e x t complete the parallelogram of which A and B O B are the sides Then the di agonal OD D will be t h e sum of B u t the A and the negative of B Fro 4 segment O D is parallel and equal to B A Hence B A ma be taken as the difference to the two vectors A and B This leads to t h e foll o wing rule Th e di ffer ence of two vectors whi ch are drawn from the same origin is the vec t or drawn fro m the terminus of the vecto r to be s u b tracte d to t h e te rminus of the vector from which it is s u b tracted Thus the two diagonals of the parallelogram which is constructed upon A a n d B as sides give the sum and di f ference of A and B In the foregoing parag raphs addition sub traction and scalar multiplication of vectors have been de fi ned and inte r To make the development of vector algeb ra ma t he p re t e d m a t ic a l l y e x act and systematic it wo ul d now become necessary to de mo ns tra te that these three fundamental operations foll o w the same formal laws as in the ordi nary scalar algebra a l though from the standpoint of the physical and geometrical interpretati on of vectors this may seem super uous These laws are . L . . , . . ’ . . . , y . . , . , . , , . , fl I : m (n A) I, II Il l a III , III, ’ ° m (A + B ) m A + m B, — A — B . . ADDI TI ON A ND S CALAR M P UL TI L I CA TI ON 13 is the s o -called law of association and commutation of the scalar factors in scalar multiplication Ib is the law of association for vectors in vector addi tion It states that in addi ng vectors parentheses may be ins erted at an points without altering t h e result II is the commuta t ive law of vector addition III is the di stri b utive law for scalars in scalar multipli cation III, is the distri b utive l a w for vecto rs in scalar mul tipli cat ion III is the distribu t ive law for the negative sign Th e proofs of these laws of operation depend upon those propo sitions in elementary geometry which have to deal wi t h the fi rst properties of the parallelogra m and si m ilar triangles They wil l not b e given here ; b u t it is suggested that the reader work them o u t for the sake of fi x ing the fundamental ideas of ad di tion sub traction and scalar multiplication more clearly in mind Th e res ul t of the laws may b e summed up in the statement Ia . . y . . , . . . , . , , . Th e l a ws wh ic h go ve rn m u l tip lic a tio n p o e ra t io n s i n f ve c to rs o a ddi tio n , s u btra c ti o n , a nd sc a la r ide n t i c a l wi th tho s e go ve rn in g th e s e l a r a lge bra a re o rdin a ry s c a . is precisely thi s identity of formal laws which justi fi es the e x tension of the use of the familiar signs and of ari t hmetic to the alge b ra of vectors and it is also this which ensures the correctn ess of results ob ta ined b y operat ing with t hose signs in t h e usual manner On e caution only need b e mentioned Scalars and vectors are en t irely di fferent Fo r t his reason they can never b e equated sorts of quantit to each other e x cept perhaps in the trivial case where each is Fo r the same reason they are not to be added to gether z ero So long as t his is borne in mind no difii c ul t y n e ed b e antici p ated from dealing wi th vectors much as if they were scala rs Thus fr om equations in which the vecto s enter lin early with It 0 . y . . . . r . VECTOR A NA L YSIS 14 r y scalar coeffi cients unknown vecto s ma be el iminated or found b y solution in the same way and wi th the same limita tions a s in ordinary algebra ; for the eliminations and solu tions depend solely on the scalar coe ffi cients of the equations and not at all on what the variab les represent If for instance . then A t hree as B , , C, or D y ma be e xpressed in terms — 1 — Cl vector equations such as An d two 2 A+ 3 B =F and yield by the usual processes the solution s A= 3 E B = 3 E Co mp o n e n ts — 4 F Z E ~ . f Ve c to rs o are said to be collinear when they are parallel to the same line ; coplanar when parallel to the same plane Two or more vecto rs to which no line can be dra wn parallel are said to be non -collinear Three or more vectors to which no plane can be drawn parallel said to be non -coplanar Obviously any two vecto s are 00 planar An y vector b collinear with a may be e xpressed as the product of a and a positive or negative scalar which is the ratio of the magnitude of b to that of a Th e sign is posit ive when b and a have the same directi on ; negative when they have op p osite directions If then 0A a the vector r drawn Def in itio n : Vectors , . . r . . . , . , ADDI TI ON AN D S y CALAR M P UL TI LI CA TI ON from the origin 0 to an point of the lin e either direction is r x a OA 15 produced in 1 ( ) . i be a variable scalar parameter th s equation may there fore b e regarded as the ( vector) equation of all points in the line O A et now B b e any point not upon the line 0A or that lin e produced in either direction ( Fig et O B b Th e vector b is surely not of t h e form se a D raw thr ough B F c 5 a line pa ra llel to 0A and let R be a n y point upon it Th e vector B B is collinear with a and is consequentl y e xp essible as m a Hence the vector drawn from O to R is If x L . L . . . r r . . . . 0R = OB + B R r =b + x a This equation may be regarded as the ( vector) equation of all the points in the line whi ch is parallel to a and of which B is one point 14 ] An y vector r coplanar with two non -collinear vecto rs a and b may be resolved into two components parallel to a and b respectively This resolution may b e accomplished b y construc t ing the pa r 6 of wh i ch the sides are a ll e l o ra m ( Fi ) g g parallel to a and b and of which the di agonal is r O f these components one is m a ; the other y b 5c and y are re s p e c t ive ly the scalar ratios ( taken with the proper sign) of the lengt hs of these components to the lengths of a and b Hence r x a yb . . . . . , . y is a typical form for an vector 00plan ar with a and b If several vecto s r 1 r 2 r 3 may b e e xpressed in this form as r , , . VECTOR A NA L YSIS 16 their sum r is then Th is the well -kno wn theorem that the compon ents of a sum of vectors are the sums of the components of those vecto rs If the vector r is z ero each of its components must th e one vector e uation r O is Conse uentl b e z ero eq u ivalent to the two scalar equations is q . . y q t = 0 3 . y vector r in space ma be resolved into three components para llel to any t hree given non -coplanar vec t o rs et the vectors b e a b and c Th e resoluti on may then b e a c c o m p li s h e d b y constructing the parallelopiped ( Fig 7 ) of which t h e edges are parallel to a b and c and of which the di agonal is r Thi s par a l l e l o p ip e d m a y b e > a i drawn easily b pass ng F c 7 y three planes parallel re s p e c t ive l y to a and b b and c c and a through the origin 0 of t h e vector r ; and a similar set of t hree planes through its t erminus B These s ix planes w then b e parallel in pairs An y L . , , . . , \ . r . . , . , ill , VE C TOR 18 A = fl, r x a r NAL YSIS + gb + z q ’ r cc 9: ' g , y z z ’ , — ’ — cc x O, z y — y ’ z w . fl b O, z ’ - z z w . z this wo ul d not be true if a b and 0 were coplanar In t hat case one of the th ree vectors could be e xpressed in terms of the other two as B ut =m c sc a ’ n r a + yb + ’ + y b + ’ [( z i + h n c z " b, c ) [a w e m x a . , , ’ ( (m z r c — a, i ss b = 0 r Hence the indiv dual componen t s of a and b ( supposed difi e re n t ) are z ero m + m e x + m z ence H i i . in the d rect ons ’ r . ’ y ’ l this b no means necessitates cc y z to be equa re s p e c In a similar manner if a and b were col t ive ly to x y z linear it is impossible to infer that their coe ffi cients vanish in d viduall Th e theorem ma perhaps be stated as follows : B ut , ’ , i ' , . y y . In or ' t wo c a se two e qu a co rre sp o n di n g s c a y t ru e c op la na r ( Ar ts . l non-collinear c e s s a r il 18 la r ve c to rs a re e xp re s s e d ve c to rs , fi coe h e w o t t f i This . , or c ie n t s ve c to rs th re e a re be i n te rm s f o non-coplanar e qu a l B . c o lli n e a r ut ve c to rs , this is th e not n e th e th re e ve c to rs , or ; o n e ve c to r , principle will be used in the appli cations et Th e Th re e Un i t Ve c to rs i, j , k . the foregoing paragraphs the method of e xpress ing vecto rs in term s of three given non-coplanar ones has been explained Th e simplest set of three such vecto s is the rect In . r recta ngular syst em may however be either of t wo very distinct 1 — t pes In one case ( Fig 8 fi rst part) the Z axis lies upon that side of the X Y plane on which rotati on through a right angle from the X-axi s to the Y-ax is appears c o u n te rc l o c kwis e or p o si t ive a ccordin g to the convention adopted in Trigo n o m e If the X tr Thi s relation may be stated in another form ax is be di rected to the right and the Y-ax is vertically the ax is will be d irected toward the observer O r if th e X Zaxis point toward the observer and the Y-ax is to the right St the Z -axis will point upwar another method of state y . . y , . . , . d ill . , L ft h - a nde d e FIG 8 . l y . ment is common in mathematica ph sics and engineering If a right -handed s c re w b e turn ed from th e X-ax is to the Y axis it will advance along the ( p o s i tive ) Z —a sc is Such a s s tem of ax es is call ed right-ha nded positive or c o u n t e rc l o c k 2 wise It is easy to see that the Y ax is li es upon that side of the Z X-plane on which rotation from the Z -ax is to the X ax s is coun tercl o ckw se ; and the X-ax is upon that side of t Th X Y B y th X Y Z — x i t h p it iv h lf ft h t x i i t h pl wh i h pl th X t h pl 0 t i d Y- x i i . y . , , - . i 1 i m i ig d d y m i l y il bl i m b fi fi g d g ig d m b fig d f m p l m p di l if dfi g b d p lm ig g l fi fi g ig d d y m f m d b fig k d m b fi fi g dfi g ea ns A o fthe fi rs t e se c o n , er be n er s n n s t re t c he be er . s c h t-h a n rs t or e os e con a ns e e s o ut ro e wa r y th e t he the n a fin wh e rs c e r o ft he r a a ta t en r en a h is a . c . an in t h e ea n s a va to a . e c o ns s t s Ift h e t h u . to e a c h es e . a ne 2 e , wa h t ha n cu a r ht s s s, a an pe r a a o a nd o ne st e e r, a n d se c o n e n t o ve r t o is s te a a ne e c o n ve n e n t r thu n se c o n han or e a ne 3 m , o rs t o r e r th u a nd the t h e r, a n d n , e r, a r rs t n ht e r, VEC TOR A NAL YSIS 20 the YZ -plane on which rotation from the Y axis to the Z a xis is counterclockwise Thus it appears that t h e relation between the three ax es is perfec t ly symmetri cal so long as the same cyclic order X YZ X Y is o b served If a righ t -handed screw is turned from one axis toward the ne xt it advances along the third In t h e other case ( Fig 8 second part ) the Z -a x is lies upon that side of the X Y-plane on which rotation through a right angle from the X-axis to the Y-axis a ppears c lo c kwis e or n e g a ti ve Th e Y-a xis then lies upon that side of the Z X plane on whi ch rotation from the Z -a xis to the X a x is appears clockwise and a similar s t a tement may be made concerning the X-axis in its relation to the YZ —plane In this case too the relation between the three a xes is symm etrical s o long as the same cyclic order X YZ X Y is preserved b u t it is just the Opposite of that in the former case Ifa left-handed screw is t urned from one a xis toward the nex t it advances along the third Hence this system is cal l ed left-handed negative 1 or clockwise Th e two sys t ems are not superposable They are sym metric O n e is t h e image of the other as seen in a mirror If t h e X and Y-ax es of the two difi e re n t syste m s be superimposed the Z -ax es will point in opposite di rections Thus one system may be obtained from the other by reversing the direction of o n e of the ax es A little thought will show that if two of the ax es be reversed in direction the system will not b e altered but if all th re e b e so reve sed it will b e Whi ch of the two syste m s be used m a t t e rs little B u t in a s m u c h as the formul ae of geometr and mech a nics dif f er y slightly in the matter of sign it is advisab le to settle once for all which shall be adopte d In th s book the right-handed or counterc lock wise system will be invariab ly employed - . . . , . - . - . , , . , . , . . . . , . . r , . , , . 1 o ne m mg le ft-h a n d e d s ys t e wa s fo r e d b y t h e ri A ma y b e ht . m . i fo r e d by t h e l e ft . ha n d j us t a s i a r g h t -ha n d e d fi ADDI TI ON A ND S CALAR j L TI P LI CA TI ON 21 MU three letters i k will be reserved to de note three vectors o f un it length drawn respe ctively in the directions of the X Y and Z ax e s of a right-handed re c ta n gular system In terms of these vectors any vecto r ma be e x pressed as De Th e n i ti o n , , y . , 6 ( ) coe ffi cients x y z are the ordi nary Cartesian coordinates of the term inus of r if its origin be situated at the origin of co ordinates Th e components of r parallel to the X Y a n d Z -a x es are respec t ively Th e , , . t a y fi z k . j rotations about i fro m jto k about from k to i and about k from i to are all pos itive B y means of these vectors i k such a correspondence is established between vector analysis and the analysis in Ca r t e s ia n co ordi nates that it becomes possible to pass a t will from either one to the other Th ere is nothi ng c o n t ra dic tor between them O n the contrar it is often desirab le or even neces sar to translate t h e fo rm ul m ob tained b y vector methods into Cartesian co ordinates for the sake of comparing t hem with results a lready kno wn and it is still more frequently convenient to pass from Cartesian analysis to vectors both on accoun t of the b revit y thereby obta ined and b ecause the vector e xpressions show forth the intrinsic meaning of the fo rmul ae Th e j , , y j . , y . y , . . Ap p lic a tio n s Problems in plane geometry may frequently be solved easily b y vector methods A n y two no n -c ollinear vecto s in the plane may be taken as the fundamen t al ones in terms of which all others in that plan e may be e x pressed Th e origin may also be selected at pleasure O ften it is possib le t o r . . . VE C TOR A NAL YSIS 22 make such an advanta geous choice of the origin and fun da mental vectors that the analytic work of solution is ma teriall y simplifi ed Th e adapta b ility of the vector me thod s about the same as that of ob lique Cartesian co ord nate s with difi e r x es upon the two a en a es c s l t Th e line which joins one vert e x of a parallelo E xa mfi dg ; d l r a m mid e point of an opposite side the diag fi f h e t n s e t s c g onal (Fig et A B OD be the parallelogram B E the lin e joining t h e verte x B to the middle point E of the side A D R the point in which this line cuts the diagonal A C To show A B is one th rd of A O Ch o o s e A as origin A B and A D as t h e F G 9 two fun damenta l vectors 8 and T Then A O is the sum of S and T Le t E B ) To show i . i . . ’ ’ L ’ ' . , i , . . I . , . . M / . . l where a; is the rati o of E R to EB R= y S ( An d ’ where y r — an l unkn own sca ar T), is the scalar atio of A g to AO q l to be sho wn e ua to Hence ; T+ n (S Hence e quating correspondi g coef cient , n fi s . r A ( t . ADDI TI ON From AN D S CALAR M UL TIP LI CA TI ON 23 which i i as a: s also 1 the line E B must be t r secte d a s well as the diagon al A O E aea mp le If through an p o int within a triangle lines D ? < be drawn parallel to th e sides the sum of the ratios of these lines to their correspo ndi ng S1 de s is 2 et A B C be the trian gle R the point within it Choos e A as origin A B and A O as the two fundamenta l vectors 8 an d T et Inasmuch y . ‘ a / fl L . ’ , . L . , ( ) a i m S is y A the fract on of A B which is cut o ff b the line th rough R parallel to A O Th e remainder of A B mus t be the frac t on 1 m 8 Conse uentl b simi ar t ian gles the ratio of the l ine parallel to A O to the line A C itself is ( 1 m ) Similarly the ratio of the line parall el to A B to the l ine A B itse f is ( 1 n ) N e x t express B in te rms of S and T S the th rd side of the tri angle Evidently from ( a ) i . . q yy l r ’ . l i . . . — S) . y Hence ( m n ) S is the fraction of A B whi ch is cut o ffb the line through R parallel to B 0 Conse uently by similar tri angles the ratio o f th s n e to B O itse f is ( m n ) A dding the three a t ios i li r l . q . (1 and the the o rem proved E xa mp le 3 If from any poin t withi n a p a ra ll e l o gra m lines be drawn parallel to th e sides the diagonals of the pa allelo grams thus formed intersect upon the di agonal of the given parallelogram et A B CD be a parallelogram R a o int within it KM an L N two lines to A B and e l resp ect vel f F 1s . -f ‘ r , L d . ’ , p i y , VE C TOR A NAL YSIS 24 the point s A D, D B O, C_ re s pe c ti l lelograms KE N D a n d of the two p araon A C T Ch o o s e A as origin A B and E as the two funda mental vectors 8 and T et LM fl a v or Ann L . c p c w 4 . R flu .. I 3 . , AR m S n T, and let P be the point of inte rsection of KN wi th L M Then — n P =A P =A LM = K+ x . )T , KM I ( P = A P = A L + y LM E quat ng i c o e fii c ie n t s , n By soluti on , + se l ( n so m + n — — n ) -r 1 m m + n — 1 either of these solutions in the e x pression m + n which shows that P is collinear with A O Problems in three dimensional geometr ma be solved in essentially the same manner as those in two dim e n sions In this case there are three fundamental vec t ors in terms of whi ch all others can be e xp essed Th e method of solution is ana logous to that in the simpler case Two y y . . r . . VE C TOR ANA LYSIS 26 manner ’ A B z x 0+ y /2 D 2 — B ) . — B ) O = 1 + k2 x 2 (l = k 2 m, = k y2 Henc In k, e P B ’ B B ’ n 1 . l — kr l a i y the same wa y t ma be shown tha t P C ' " c o ” D d DD ri Addi ng t h e ’ F ‘ i four at os the res ul t s 1 r i a lin e which passes th ough a g ve n point and cuts two given lines in space et the two lines be fi x ed respec t ively by two poin ts A and B O and D on each et 0 be the g ven point Choose it as orig n and let E xa mp le 2 To fi n d : L . , i . A = OA , An y An y po int P o fA po int Q of If th e points a re c o llin e a r . . P B CD L B = OB y i , . D = 0D d . ma be expresse as — A) — O) i may likew se be written and Q li e in the That is P = z sa i r me l n e th ough Q . 0, d P an Q A DDI TI ON AND ALAR SC M P UL TI LI CA TI ON 27 efore it is possib le to equate coeffi cients one of the vectors must be expressed in te rms of the other three et Then P = A + ze ( B — A) B L . Henc l e — x ac H enc y l, m y , z z m e a : l + m 9 1 ituting in P and Subst Q ither of these may be taken as de ing a fin E and cutting AB and dr a wn OD ’ . Vec to r Re la tio ns i n dep e n de n t of th e Origin 20 ] m n . Choose y divide a line To E xa mp le 1 F i ( g L 0 . P = 0P = 0A + That is , i in a g ven . i r AB ry an arbit ar po int 0 as origin et OA A and O B B To fi n d the vecto r P O P of which the terminus P d vides A B in the ati o m n . from m m + A B =A+ Fro . 10 . O VE C TOR ANA L YSIS 28 r component s of P parallel to A and B a re in inve se ratio t o the segm ents A P and P B into which the line A B is di vided b y t h e point P If it should so happen t hat P divided t h e line A B e x ternally the ratio A P / P B would be nega t ive and t h e signs of m and n would be oppos ite b u t t h e formula would hold without change if this difference of sign in m and n be taken into account To fi n d the point of intersection of the me di ans E xa mp le 2 of a triangle et A B C be the given Choose the origin 0 at ran dom et OA A O B B and O C G et t riangle C b e respectively the middle po ints of t h e sides opposi t e t h e vertices A B 0 et M be the point of intersection of the medians and M OM the vec t o r drawn t o it Then Th e . , , , . . L . , . . , , , L L ’ L . . (B — A) — B 2 (C ) 2 Assuming that 0 has been chosen outside of the plane of the triangle so that A B C are non-COp l a n a r corresponding c o e ffi eleu t s ma be e uated y q , , , . x = l — y, Hence Hence 5 (A B C) . A DDI TI ON AND ALAR M SC P UL TI LI CA TI ON 29 vector drawn to the median point of a triangle is equal to one third of the sum of the vecto rs drawn to t he vertices In the pro b lems of which t h e solution has j u st b een given the orig in co ul d b e chosen arb itrarily and the result is in dependent o f that choice Hence it is even possible to dis re gard the origin entirely and replace the vectors A B 0 etc Thus the points themselves b y their termin i A B C etc become the subjects of analysis and the formul ae read Th e . . , , , , , . , . , n A + m B m + n M y This y is t pical of a whole class of problems soluble b vector methods In fact a n y p u re ly ge o m e tri c relati on b etween t h e difi e re n t parts of a fi gure must neces sarily be independent of the origin assumed for the analytic demonstration In some cases such as those in Ar t s 1 8 1 9 the position of th e origin may be specializ ed with regard to some crucial point of the fi gure so as to facilitate the computation ; b u t in many other cases the generality obtained b y leaving the ori gin u n Speciali z ed and undeterm ined leads to a s y mm e try w hi ch renders the results just as eas y to compute and more easy to rememb er Th e o rem : Th e necessar and su f fi cient condition that a vector equation represent a relation independent of the origin is that the sum of the scalar coeffi cients of the vectors on one side of the sign of equalit is equal to the sum of the coe ffi cients of the vectors upon the other side Or if all the terms of a vector equation b e tra nsposed to one side leaving z ero on the other the sum of the scalar coe ffi cient s must be z ero et the e uation wri tte n in the latte r form he . . , . , , y . y . , L . q 0 . VE C TOR ANA L YSIS 30 C ha nge the origin from O to O b y addi ng a constan t vector Th e equation R O O to each of the vectors A B C D then becomes ’ ’ , , , o this is to be independent of the origin the coeffi cient of mus t vanish Hence If B . That thi s condition is ful fi lled in the two e xamples cited is obvious . If m necessary and su ffi cient condition that two vectors sa t isfy an equation in which the sum of the scalar coeffi cients i s z ero is that the vectors be equal in magnitude and in direction Th e , , . Firs t l e t and a A+ bB = O a + b 0 z . is of course assumed that not both the coeffi cients a and b vanish If they did the equation would mean nothing Su b s t it u t e the value of a obtained from the second e uation into It q . t h e fi rs t Hence . A B . . A DDI TI ON Secon dl q y if A an ALA R q 31 i are e ual in magni tude and d recti on d B e uati on th e M UL TIP LI CA TI ON AN D SC A 0 B subsists between them Th e sum of the coeffi cient s is z ero fi cient condition tha t th ree vecto s Th e necessar and s u f sa tisfy an e uation in wh ich the sum of the scalar coeffi cients is z ero is that when drawn from a common origin the te rmi ) nate in th e same s t raight line q y r . . y , , E mH % a A + bB + md a + b + C= 0 c c =fi i q dq all the coeffi cients a b c van sh or the e uations wo ul d b e mean ingless et c be a non-vanis hing coe ffi cient Substitute the value of a obta ned from the secon e uation No t L . in t o t h e fi rs t , , , i . . c (C — i B i ) . d Hence the vector wh ch jo ns the e xtremiti es of C an A is collinear w th that which joins the e x tremities of A and B Hence those three points A B C lie on a line Secondl y suppose three vectors A O A B O B C 0 0 drawn from the same origin 0 termin ate in a stra ight line Then the vecto s i . ’ , . , , r are co llinear i , . A B =B . — A a nd q A O= C— A Hence the e uati on id i subs sts Th e sum of the coeffi cients on the t wo s es s t h e same Th e necessar and su f fi cient condition that an e uati on in which the sum of the scalar coe ffi cients s z ero subs ist . . 1 Ha m y mm igi V e c t o rs wh i c h ha ve a c o ” “ ilt o n l a min a c o lline a r . on o r n a nd mi te r na te q i in o ne , , lin e a re c a ll e d by VECTOR 32 A NAL YSIS between four vectors is that if drawn from a common 1 they terminate in one plane , . Firs t l e t and L et of a be a non-van ishi ng coe ffi cient Sub sti t u te obtained from the last equation into the fi rst d t he . . (D d — C) . line A D is coplanar with A B and A 0 Hence all four termini A B C D of A B C D li e in one pla n e Secondl suppose that the termini of A B C D do lie in one plane — A an d A B Then A D = D B — A a re c c planar vectors O n e o f them may b e ex pressed in terms of the other two This leads to the equation Th e y . , , , , , , . , , . , z , . . l where — (B - A) 0, and n are certain scalars Th e sum of the c o e ffi c ie n t s in th s equation is z ero B etween any fi ve vectors there e x ists one equation the sum of whose coeffi cients is z ero et A B C D E be the fi ve given vectors Form the differences l, m , L i . . . , , , . , E _ A, _B, E E _ C, E — D . of these may be e x pressed in terms of the other three — o r what amounts to the same thing there must e xist an equation be t ween them On e . 70 ( 13 Th e 1 by — 11 ) + l (n — B)+ m (E q i sum of the coeffi cients of th is e uation s z ero m Ve c t o rs wh ic h ha ve a c o H a ilto n te rmin o — c o mp la mm i gi on or n a r. " n a nd m te r in a t e in o ne . p la n e a re ca lle d VECTOR A NAL YSIS 34 r However the points E C and A lie upon the same st aight line Hence the equation which connects the vectors E C and A must be such that the sum of its coe ffi cient s is z ero n This determines x as 1 , , , . . . Hence By E — n C= D— n B = (1 — n )A . another rearrangement and similar reasoning Subt ract the fi rst equation from the second — This n )A . vector cuts B C and A G It must therefore b e a multiple of F and such a multiple that the sum of the c o e ffi c ie n t s of the equat ons whi ch conn ect B C and F o r G A and F shall b e z ero . i , , , , . Hence B + C Hence 2 d and the theorem has been proved Th e proof h a s covere considera b le space because each detail of the reasoning has In realit however the actual anal sis h a s c o n b een given sisted of just four e uations ob tained simply from the fi rst E xa mp le 2 : To determine the equations of the line and plane et the line be fi xed b y two points A and B upon it et P b e an poin t of the line Choose an arbitrar origin Th e vecto s A B and P terminate in the same ne enc e . L y q . y , , . . y r , , a Therefo r li . A + bB + p P = 0 = a + b + Q p a A + bB e a b y . . H L . i A DDI TI ON ALA R T ON 35 M UL TIP LI CA I AN D SC d fferent point s P the scalars a a n d b have different values They ma be replaced by x and y which are used more generall to represent variables Then Fo r y y . , . x A+ yB + y x L et a plane determined by th ee poin s and L et be any point of the pla e ho se an arbitrary origin vecto rs P t rmi ate in one p lane Hence be P n Th e A, B , C, an d e a As a , b, customa ry c, 0 , o n . . . A + bB + + b + a y C . A, B t r C c c i var for di fferent point s of the plane it s more to write in th eir ste ad x y z , , x , A+ yB + x + y + . z C z line whi ch joins one ver te x of a com i quadrilateral to the ntersection of two d agonals l t e e p divides the opposite sides har E xa mp le 3 Th e : i i F ( g A , B , O, D m o n ic a ll y L et be four vertices of a quadrilat ral L et meet . e AB . in a fi fth verte x E and A D meet B C in the six th vertex F 5 et the t wo di agonals A C and F G 12 To show B D inte rsect in 0 that F G intersects A B in a poin t E and CD in a point E such that the lines A B and CD are divi ded inte rnall at E and E in the same ratio as they are d vided e xte rna y b E That is to show that the cross rat os CD , L . I . . . ’ ’ y ” ' i . (A B EE ’ ) ( OD E E o ” ) y i — r . ” ll 36 VE C TOR ii Choos e A the or g n at random drawn from it to t h e points plane Henc e NAL YS S Th e . A, B r r fou vecto s A B O D terminate in , , , , C, D o ne . Separate q i y the e uat ons b transpos ing two terms d A+ C= + a a Di vi de c c A+ bB + dD C b + d In lik e m a n n e r F d A + dD a ( ( Separate bB + c b + c + d + )G a a + c > ( + )G a c c c — q d + a i C c C C — dD c e ( c — ) dD — d the e uations again and d vide a Hence c A + bB a i c C + dD + b c + d d vides A B in the ratio a : b and CD in ” divides CD n the c d B u t e uation ( a ) shows that E Hence E and E divide OD internall and c : d ra tio e x ternally in the same ratio Whi ch of the two d v sions is inte rnal and which ex ternal depends upon the relat ve signs of c and d If the have the same sign the internal point of division is E ; if opposite sign s it s E In a similar wa E and E may b e shown to div de A B harmonically Exa mp le 4 : To di scuss geometric nets B y a geometric net in a plane is meant a fi gu e composed of points and straight lines obtained in the follow ng manner Start w th a certain number of oin t s all of wh ch lie in one E q . ” ii i . . . y i ’ , i y ” . . . i i y p i r i . A DDI TI ON A ND ALA R M SC P UL TI LI CA TI ON 37 plane Draw all the lines joining these points in pairs These lines will intersect each other in a number of points N e x t draw all the lines whi ch connect these points in pairs Thi s second set of lines wil l determine a still greate r number of points which may in turn be joined in pairs and so on At each step Th e co n struction m a y be kept up inde fi ni tel the numbe r of points and li nes in the fi gure increases Prob ab ly the mos t interesting case of a plane geometric net is that in which fo u r poin t s are given to commence with oining these there are s ix lines whi ch inte rsect in three J point s d fferent from the given four Three new lines ma now be drawn in the fi gure Thes e cut out s ix new points From thes e mo re lines may be obtained and so on To trea t th s net an a ytically w te down t h e e uat ons . . . . y . . . . i y . . i . l ri q i . and which subsist be tween the fo u r vectors drawn from an unde t e rm in e d origin to the fo u r given points From these t is poss ible to obta n i a A + bB a a a c + c C= dD c C 9 + d bB + dD b + d A + dD a c + b A+ a li i . bB + c b + c + d C sp tting the equations into tw o parts and dividin g N ext four vectors such as A D E F ma be chosen and the e ua tion the sum of whose coe ffi cient s is z ero ma y be determ ned This would be by , i q , , treating th s e uati on as ( c ) be obtained By . y wa s t q i . reated ne w . p int may o s VECTOR ANAL YSIS 38 —d A + dD _ 2 + d a a + b + b + d dD+ a c 9 c +d e + b) E (a + b + d E quations between other se t s of four vecto rs selected from A B C D E F G may be found ; and from these more points obt ained Th e process of fi nding more points goes forward inde fi nitely A ful ler account of geometric nets ma y be found in Hamilton s E le m e n ts of Qu a te rn io n s B ook I As regards geometric nets in space jus t a word may be said Five points are given From these new points may be obta ined by fi n di ng the intersections of planes passed thr ough sets of three of t h e given points w th lin es connecting the remaining pairs Th e cons truction may then be carried for ward with the points thus ob tain ed Th e anal tic treatment is similar to th at in the case of plane nets There are fi ve vectors drawn from an undetermined origin to the given fi ve points B etween these vecto s there e x ists an equation the sum of whose coe ffi cients s z ero This equation may be separated into parts as before and the new point s ma thus be ob tained , , , , , , . . ’ ” , . . . i y . . . r i . y . . and a A + bB a a + b A+ a c c C + dD + c C + d + b + d + E e E e bB + dD + + b e e are two of the poin t s and others may be found in the same wa y N ets in sp a ce are als o s cus sed by Hami lton lo c c it . di , . . ADDI TI ON AND S CA LAR Ce n te rs Th e M P UL TI L I CA TI ON 39 G r a vi ty f O y f gr a vi ty of a s stem of particles may c e n te r o foun d very easily b y vec t or methods Th e two laws of ph sics which will b e assumed are the following : Th e center of gravity of t wo mas ses ( considered as situated at points ) lies on the lin e connectin g the two mas ses and divides it into two segment s which a re inversel pro portional to the masses at the ex trem iti es In fi n ding the center of gravity of two systems of masses each system ma b e replaced by a sin gle m a ss equal in magnitude to the sum of the masses in the system and situated at the center of gravity of the system G iven two masses a and b situated at two points A and B Their cente r of gravit G is given b be y . y . y . y y . A+ bB a 9 + b a y where the vecto rs are referred to an origin whatsoever This follows imme di ately from law 1 and the formula ( 7 ) for division of a lin e in a given ratio Th e center of gravit of three masses a b c situated at the three points A B C may be found b y means of la w 2 Th e masses a and b ma be considered as equivalent to a single mass a b situated at th e point y , , . , y A + bB + b a G (a a Hence , . a Then . a + b + A+ bB + a b c c c C 40 E videntl a , be b, c, y VE C TOR ANAL YSIS th e d, found in a y cente r of gravit of any number of masse s si tuated at the points A B C D si milar manner Th e esult is r . , , , , r iy lines which join the center of g a v t of a triangle to the vert ces divide it into three tri angles whi ch a re proportion a l to the masses at t h e o p F i i vertices et A B s t e C o g p ( be the vertices of a tri angle weighted with masses a b c et C be the cen ter of gravi t oin A B C t o G and J Fro 1 3 produce the lines until they in te rsect the opposite sides in A B C res ectivel To show th at the areas Th e Th e o re m 1 i L . y . . ’ ’ last proporti on between A B from compounding the fi rst three the demonstration Th e , L . , , . p ’ , , , , y , . c comes and a b It is however us eful n C , . i , . Henc In AB C A A ’ CB C G A ’ AB C e A C ’ G A a + b+ G B C a s i mi lar man ner CA ’ b + c a c a B CA a G CA CA B r + CA ’ GAB + b + c 6 a + b + c c Hence the p oportion is proved The o re m 2 : Th e lines wh ch join the center of gravi t of a te t rahedron to the vertices d vide the tetrahedron into four i . i y VECTOR ANA L YSIS 42 may therefore be looked upon as co ordi nate s of the points P ins ide of the triangle A B C To each set there corresponds a de fi nite point P and t o each po in t P t here corresponds an in fi nite n umber of sets of quantities which however do not di ffer from one another e xcept for a factor of proportionality To obtain the points P of the plane A B C which lie o u ts ide of the triangle A B C one may resort to the conception of negative weights or masses Th e cente r of gravity of the masses 2 and 1 situated at t h e points A and B respectively would b e a point G dividi ng the li ne A B e xte r n a lly in the ratio 1 2 That is a , b, ' c . , , . . . y y point of t h e line A B produced ma be represented b a suitab le set of mas ses a b w hi ch difi e r in sign Similarly any point P of the plane A B C may be represented by a s u itab le set of m a sses a b c of which one will di ffer in sign from the other two if the point P lies outs ide of the triangle Inasmuch as only the ratios of a b and c are im AB C portant t wo of the quantities may always be ta k en positi ve Th e idea of employi ng the masses situated at the vertices as coordinates of t h e cente r of gravity is due to M ob ius and was publis hed b y him in his book entitled B a ry c e n tris c h e Ca lc i tl in 1 8 2 6 This may be fairl regarded as the start ing point of modern analyt ic geometr Th e concept ion of negative masses which have no e xistence in nature may be avoided b replacing t h e masses at the vertices b y the areas of the triangles G B C G CA and G A B to whi ch they are proportional Th e co ordi nates of a poin t P would then b e three numbe s proportional to the areas of the three triangles of which P is the common ver t e x ; and the sides of a given triangle A B C the bases Th e sign of these areas is determined b the follo wing de fi nition An y , , . , , . , . ” , y . “ y . y , , r . y , . . fi A DDI TI ON AND S CA LAR M P UL TI LI CA TI ON 43 area A B C of a tri angle is said to be positive when the vertices A B C fol low each other in the positi ve or counterclockwise di recti on upon the circle de scri bed thr ough them Th e area is said to be negative when t h e poin t s foll ow in th e negative or clockwise direction Cyclic permutation of the lette s t herefore does not alter the sign of the area De Th e n i tio n : , , . r . . A B C = B CA = C A B . rl Interchan ge of two letters which amounts to a reve sa of the c cli c order changes the sign y . A CB = B A C = CB A IfP be an y — AB C . point withi n the triangle the equation P A B + P B C + P CA = A B C mus t hold Th e same will also hold if P be outside of the triangle provided the s ign s of the are as be t aken into con sideration Th e areas or th ree uantities proportiona l to them ma be regarded a s coo rdi nates of the point P Th e e x tension of the idea of ba ry c e n tric co ordin ates to space is imme di ate Th e four points A B C D situate d at t h e verti ces of a tetrahedr on are weighted wi th mass a b c d respectively Th e cent er of gravity G is represented b y these uan tities or four others proportional to them To obta in points outside of the tetrahedron negative masses may be emplo ed O r in the light of theorem 2 page 4 0 the masses may be replaced by the four tetrahedra which are proportional to them Then the idea of negative vol umes ta k es the place of that of negative weights As thi s idea is of considerab le importa nce later a brief treatment of it here may not be out of place Th e volume A B CD of a tetrahedron is said Defin i tio n to be positive when the triangle A B C appe a s positive t o . q y . . ” “ , . , , , q , , . . y . , . . , . : r , 44 VE C TOR y A NAL YSIS i i the e e situated at the point D Th e volume s n e gat ve if the area of the tri angle appear negative To make the discussion of the sig n s of the vario u s tetrahedra perfectl clear it is a lmo s t n e c e ss a ry to have a solid mode A p la n e draw ng is scarcely s u ffi cient It is di ffi cult to see from it wh ch triangles appear posi t ive and which negative Th e following relat ons wi b e seen t o hold if a model be ex amined Th e in terchange of two letters in the te trahedron A B CD changes the sign . . y l i . i . i . ll . . A C B D = C B A D = B A C D = D B CA = A D CB = A B D C = — A B CD ' . sign of the tetra hedron for any given one of t h e p o s sible twent -four arran gements of the letters ma be obtained by reducing that arrangement to the order A B C D b means of a number of successive inte rchanges of t wo ette s If the number of interchanges s even the sign is the sam e as that of A B CD ; if odd opposite Thus Th e y y l r i , CA D B If P — . . C A B D = + A CB D is any point in ide of quat on y s — A B CD the tetrahedron . A B CD t h e i e A R OP — B CD P + C D A P ill i i — DA B P i z A B CD r holds good It st s true f P be w thout the tetrahed on provi ded the signs of the volumes be taken into considera tion Th e equation ma b e put into a form more s y m m e t ri cal and more easil remembered b transposing all the terms to one number Then . y . y y . A B CD + B CD P + r proportion in theorem 2 page 4 0 does not hold t u e f the signs of the te trah edra be rega rded It should read i Th e , , . A DDI TI ON AND S CALAR P UL TI L I CA TI ON M 45 the point G lies ins ide the te tr ahe dron a b c d re p re sent quantities proporti onal to the masses wh ich must be l o cated at the vertices A B C D respectivel if G s to be t h e center of gravit If G lies outside of the te trahedron the may still be regarded as masses some of wh ch are negati ve — o r perhaps better merel as four numb ers whose ratios determine the pos it on of t he poin t G In this manner a set of ba ry c e n tri c coo rdi nates is establis hed for Space Th e vector P dra wn from an indeterminate orig n t o an poin t of th e pla ne A B C is ( page 3 5) If , , y , , y , i y i . y i , “ . i . xA + yB + x + y + ring this with the e pression Co mpa z C c C y z x A + bB + a a + b + c it will seen that the quantities are in reality nothing mo re nor less than the ba y centric co rdinates of the point with respect to the tri angle like ma er from e q uation be x, y z P o r AB C . x , In nn A+ yB + z C+ wD + y + z + w x y y dr y ri a A+ bB + c C+ dD + b + c + d which expresses an vector P awn from an indete m nate origin in term s of four given vectors A B C D drawn from the same origin it ma be seen b comparison w th , r a i , , , r i y y tha t the fou quantit es x y z w a e pre cis el the bar centri c co ordi nate s of P the terminus of P with respect to the te trahedr on A B CD Thus the vector methods in wh ch “ the origin is undeterm ned and the methods of the B a ry o u l u s are pract cally c e xtens ve C a l c c e n t ric It was m e nt one be fo e a n d t m a y b e w e to epe a t he r e , , , i i d i , , . r i i i . ll r VECTOR ANAL YSIS 46 that the origin may be left wholly out of consideration and Th e vector e uat ons t h e vectors replaced b y their termini then become point e uations q x q i . A + yB + z C + wD This step b rings in the points themselves as the objects of alysis and leads still nearer to the B a ry c e n tri s c he Ca lc ul of M ob iu s and the A u s de h n u n gs le h re of G rassmann ” “ . fi Us e The f Ve c to rs O to de n o t e A re a y s area l ing in one plane M N and b ounded b y a continuous curve P QR which nowhere cuts itself is said to appear positive from the point 0 when the lett ers P QR follow each other in the counterclockwise or positive order negative ; N when the follow in the nega tive or clockwise order De An n i t io n y Fi ( g It , i . is ev dent that an area can have no dete rmined sign r se n l e but y in reference o Fro 1 4 p to that directi on in which its boundar is supposed to be traced and to some point 0 out side o f its plane Fo r the area P R Q is negative relative to P Q R and an area viewed from O is negative relative to the same area viewed from a point 0 upon the side of the plane opposite to O A circle l ing in the X Y-plane and desc ri b ed in the positive trigonometric order appears positive from ever point on that side of the plane on which the positive Z -ax is lies but negative from all points on the side upon whi ch , , . y . . , y ’ y ADDI TI ON AND S CALAR M P UL TI LI CA TI ON 47 the negative Z ax is lies Fo r this reason the point of view and the di rection of description of the bo undary must be kept clearly in mind A nother method of stating the de fi nition is as follows : If a person walking upon a plane traces out a closed curve the area enclosed is said to b e positive if it lies upon his left hand side negative if upon hi s right It is clear that if t wo persons b e considered to trace out together the same curve b y walking upon opposite sides of t h e plane the area enclosed will lie upon the right hand of one and the left hand of the other To one it will consequen t ly appear positive ; to the other negative That side of the plane upon whi ch the area seems positive is cal led the positive side ; the side upon which it appears negative the negative side This idea is familiar to students of electricit and magnetism If an electric current flo w around a closed plane cu rve the l ines of magnetic force through the circ ui t pass from the negative to the positiv e side of the plane A positive magnetic pole placed upon the positive side of the plane w l be repelled b the circuit A plane area may b e looked upon as possessing more than positive or negati ve magnitude It may be considered to possess direction namely the direction of the normal to t h e posi t ive side of the plane in whi ch it lies Hence a plane area is a vector quantity Th e follow ng theore m s concerning areas when looked upon as vectors are important If a plane area be denoted by a vector whose The o re m 1 magn itude is the numerical value of that area and whose direction s the normal upon the positive side of the plane then the orthogonal projection Of that area upon a plane wi ll be represented b y the component of that vector in the direction normal to the plane of projection (Fig et the area A lie in the plane M N et it be projected ’ ’ o thogonall upon the plane M N et M N and M N inter - . . , . , . . , y , . . il . y . . , , i . . . i r L , y ’ . L . L . ’ VECTOR A NAL YSIS 48 l sect in the lin e l and let the di edra angle between these two planes b e x C onsider fi rst a rectangle P OR S in M N whose sides P O R S a n d QR SP are respectively pa rallel and perpendicular to the line I This will project into a ’ in M N Th e sides P Q and R S rectangle ’ will b e equ al to P O and R S ; but the sides Q R and S P will be equal to OR and SP multipl ied by the cosine of x the an gle between the planes C onsequently the rectangl e . , , , . ’ ’ ’ ’ ’ ’ ’ . ’ , . QR S cos P FIG Hence i . 15 x. . i y rectangles of wh ch the sides are respect vel parallel and perpendi cular to l the line of intersection of the two planes project in t o e ctangles whose sides are likewise espectively pa allel and perpendicul ar to l and wh o se a re a is r , , e qu a l t o th e c o s i ne f th e o r r a re a a n le g From f o t he , o rigi n a re c ta n gle s l be twe e n the p la n es. m u ltip l ie d by t he i y y this it follows that a n y area A is projected nt o an area which is equal to the g ven a ea mul tiplied b the cosine of t h e angle between the planes Fo r an area A ma be di vide d up into a large number of small rectangles by drawing a series of l ines n M N pa a e an perpend c u la r to the lin e I i r . i r ll l d y i . VE C TOR ANA L YSIS 50 Th e The o re m 2 y vector which represen t s a c lo s e d pol hedra l surface is z ero This may be proved by means of certain co n siderations of hydrostatics Suppos e the polyhedron drawn in a b od of fluid assumed to be free from all e x ternal forces gravi t y in cluded 1 Th e fluid is in equil ib rium under its own internal pressures Th e portion of the fluid b ounded b y the closed surface moves neither one way nor the o t her Upon each face of the surface the uid e x er t s a de fi n ite force p rop o r tio n a l to the a re a of the face and normal to it Th e resultant of all these forces must be z ero as the fluid is in equ ilibrium Hence the sum of all t h e vector are a s in the closed surface is z ero Th e proof may be given in a purely geometric manner Consider the orthogonal projection of the closed surface upon any plane This consists of a do u ble area Th e part of the surface farthest from the plane projects in to positive ar ea ; the part nearest the plane into negative area Thus the s urface projects into a certain portion of the plane which is covered twice once with positive area and once with negative These cancel e a ch other Hence the total projection of a closed surface upon a plane ( if taken with regard to sign) is z ero B u t b y theorem 1 the projection of an area upon a plane is equal to the component of the vec t or representing t hat area in the di rection pe rpendicular to tha t plane Hence the vec t or whi ch represents a closed surface has no component along the l ine perpen di cular to the plane of projection This however was a n y plan e wha tsoever Hence the vector is z ero Th e theorem has been proved for the case in which the closed surface consists of planes In case that surface b e . y . , . . fl . . , . . . . . , . , . . . . , . , . . . Suc h a s ta f i rs mg m i d t o a ll pra c t ic a l p u rp o se s in t h e c a s e o f a po lyh e dro n s us pe n d e d in t h e a t o s ph e re a n d c o n s e q u e n t ly s u bj e c te d t o a t o s Th e fo rc e o f ra vi t y a c t s b ut is c o un t e rba la n c e d by t h e t e n s i o n p h e ric p re s s u re in t h e s us pe n din st ri n 1 t e o fa f a . g g . is re a l z e A DDI TI ON A N D S CA LA R M P UL TI L I CA TI ON 51 curved it may be regarded as the li m it of a polyhedral surfa ce whose number of faces increases without limit Hence the vector which represents any closed surface polyhedral or curved is z ero If t h e surface b e not closed b u t be curved it may be represen t ed b y a vecto r just as if it were polyhedr al That vector is the li mit 1 appro a ched by the vector which represents tha t polyhedral surface of whi ch the curved surface is the limit when the num b er of faces be comes inde fin i tel grea t . . . y q SUMM AR Y or CHA PTER I vector is a uantity considered as possessing magnitude and di rection E qual vecto s possess the same magnitude and the same direction A vector is not altered by shi fting it arallel to itsel f A null or z e ro vector is one whose mag n it u de is z ero To multiply a vector by a positive scalar multiply its length b y that sca ar and leave its direction uncha nged To multiply a vector by a negative scalar mul t ip l y its length b y that scalar and reverse its direct on V ecto s add accordi ng to t h e pa rallelogram law To subtract a vector reverse it s direction and add A ddition s u b tra c tion and multipli cation of vec t ors by a scalar follow t h e same laws as addi tion subtrac t ion and multipl ication in ordinary algeb ra A vecto r may b e resolved into thr ee components parallel to any three non-coplanar vectors Thi s esolution can be accomplis hed in only one way A r . p . . l . i . r . . . , , , , r . . . components of equal vectors parallel to three given non-coplanar vecto rs are equal and conversely if the com n e n ts are equal the vect rs are equal T h three unit o e o p vectors i jk form a ri ght-handed rectan gular sys tem In Th e , , , . , 1 t he m . , li i t e xi s ts a n d is u n iqu e po lyh e dra l s urfa c e a ppro a c h e s t h e Th i s . i n d e pe n de n t c urve d s u rfa c e It is . of the m e th o d in whic h VECTOR A NAL YSIS 52 rms of them any vector may be e xpressed Cartesian co ordi nate s x y z te , , x Applications ratio m Th e . is given :n by means of the by . i + y j+ z k 6 ( ) . point which divides a l ine in a given the form ul a n A+ m B m + y (D n necessar and su ffi cient con di tion that a ve c tor equati on represent a rel ation independent of the origin is that the sum of the scalar c o e fli c ie n t s in the equation be z ero B etween any four vectors there e x ists an equation with scalar c o e fli If the sum of the c o e fli c ie n t s is z ero the vectors are c ie n ts termino-coplanar If an equation the sum of whose scalar coeffi cients is z ero ex ists b etween three vectors the are termino-collinear Th e center of gravity of a number of mas ses a b c situated a t the te rmini of the vecto s supposed to be drawn from a common o ri gin is A B C given b y t h e formul a Th e . . y . r . , , , , r i vecto r may be used to denote an area If the a ea s plane the magnit ude of the vector is equal to the magnit u de of the area and the direction of the vector is the direction of t h e normal upon the positive side of the plane Th e vector representing a c lo s ed surface is z ero A . , . . EXER CI SE S on the laws stated in Art 1 2 2 A tri an gle may be constructed whose sides are pa alle and e ua l to the med ans of an given tri an gle 1 . . D emonstrate CHA PTER I q . i y r l . . ADDI TI ON A ND S CA LAR M P UL TI LI CA TI ON 53 po ints in which the thr ee di agonals of a com 1 meet the pairs of opp site sid o e s lie three l a quadr ngle e t e p by three upon four straight lines If two tri angles are so situated in space that the three 4 points of intersecti on of corresponding sides lie on a line t hen the li nes joining t h e corresponding vertices pass through a common point and conve sely Find the mi ddle point G iven a quadrilateral in space 5 of the line whi ch joins the middle points of t h e diagonals Find the middle poin t of the li ne whi ch joins the middle points of two opposite sides Show that these t wo points are the same and coincide with the center of gravit of a system of equal masses placed at the vertices of the quadrilateral If two opposite sides of a quadr ateral in space be 6 di vided proportionally and if two quadri late rals b e formed by joining the t wo poin t s of division then the centers of gra vi ty of these two quadrilaterals lie on a li ne with the center of grav t y of the original uadrilateral B y the center of gravity is meant the center of gravity of four e ual masses pl a ced at t h e vert ices Ca n thi s theorem b e generaliz ed to the case where the masses are not equal ? Th e bisecto rs of the angles of a triangle meet in a 7 po int If the edges of a h e x a h e dro n meet four b four in three 8 points the four di agonals of the he x ahedr on m eet in a point In the special case in whi ch the he x ahedron is a parallelopiped the three points are at an in fi n ite di s tance Prove that the three straight lines through the mi ddle 9 points of the sides of any face of a te t rahedron each parallel to the straight lin e connecting a fi x ed point P with the mid dl e point of the opposite edge of the tetrahedron meet in a 3 Th e . s ix . . , r . . . . y . il . . , i q q . . . y . . . , . . , , m pl q g f p i i j i p i 1 A t h ro u h wh c h ete co o ur o n the ua g d ra n l e o n ts n o i c o n s st s o f t h e s ix s t ra i i g lli n e a r o fpa i rs o f s id es t h re e o fwhic h a re o n t s o f n t e rs e c t o n i co g li n e s wh i c h m a y b e pa ss e d Th e di a o na l s a re t h e l in e s ht . . ' VE C TOR A NAL YSIS 54 point E and that this point is such that P E passes and is b isected by the center of gravit of the t e tra h Show tha t wi tho u t e xc ep tio n there e x ists 0 10 equation with scalar coe ffi cients between an fo vectors A B C D Discuss the condi t ions imposed upon three 11 fi ve vectors if they satisfy t wo e uatio n s the sum 0 e fli c ie n t s in each of wh ch s z e o y y . , , , . . i i r q . , C H AP TER II DIR ECT A N D SKEW P ro du c ts PR OD UCTS O F VE CTOR S f Two Ve c to rs o operati ons of a ddition subtraction and scalar mul tipli cation have been de fi ned for vec t ors in the way suggested by ph sics and have been employed in a few applications It now becomes necessar to introduce two new comb inati ons of vectors These wil l be called p r o du c ts b ecause the obey the fundamental law of prod u c t s ; i e the dis tri bu tive law which states that the product of A into the sum of B and C is e ual to the sum of the products of A into B and A into 0 De n i tio n : Th e di rec t p r o du c t of t wo vectors A and B is the s c a la r quantit ob tain ed b y multiplying the product of the magnitudes of the vecto s b the c o s i n e of the an gle b e tween them Th e di rect product is denoted b wri ting the two vectors w th a dot between them as TH E , y , y . y . . . , q fi . y r y y . i A B . This is read A do t B and therefore may often be called the dot product inste ad of t h e di rect product It is al s o called the scalar product ow ng to the fact that it s value is s o a lar If A be the magnitude of A and B tha t of B then b de fi nition B = A B cos A- i . . Obv , iously the d rect product follows the commutative law i A B B A . y VECTOR A NA L YSIS 56 i ei t her vector be multiplied by a scalar the pr duct s mult iplied b y that scalar That is If o . ( x B) A) . case the two vectors A and B are colli near the angle b e tween the mbecomes z ero or one hundr ed and eighty degrees and its cosine is therefore equal to unity with the pos itive or negat ive sign ence the scalar product of two parallel vectors is num erically equal to the product of their lengths Th e sign of the product is positi ve when the di r ec t ions of the vectors a e the same negative when they a re opposite Th e product of a vector b it self s therefore e u al to the s ua e of its lengt h In H . . r y i , q A A A C onse quently if q r . 2 ( 3) y produc t of a vector b itself vanish the vector is a null vector In case the two vectors A and B are perpendicu a the angle between them b ecomes plus or m inus ninety degrees and the cosine vanishes Hence the product A B vanishes C onversely if the scalar product A B van shes then the lr . i . . , cos (A B ) 0 Hence either A or B or cos ( A B ) is z ero and eith er the vecto s are perpendic ul ar or one of them is null Thus the A B , r , fo r Th e r j i, , k is A o 0 B f vec t o rs , n e ithe r o d . scalar products of the three fun amental unit are evidently r ly more ge n e a l product If f two e n dic u la ri ty th e p e rp va n is he s , vecto s , . c o n di t io n wh ic h . a and a b b are any two uni t vecto 008 (a , b) . r s the VECTOR ANAL YSIS 58 i scalar or d rect product foll ows the ] law of multi plication Tha t is = R C R A C C A + + ) ( Th e 29 ive di stri but . . This 6 ( ) L may be proved b y means of projections et C be equal to its magn itude C multiplied b a u ni t vector 0 in its dire c tion To show y . . (A + B ) c + B -( C c ) c . is the projection of A upon 0 ; D c tha t of B upon 0 ; u t the projection of the B that of upon 0 c A B B A ) ( sum A B is equal to the sum of the projections Hence the relation ( 6) is proved B y an immedi ate genera z ation a A B r A + + + + ( P + B (1 + + B A c c o , . , li . . i scalar product may be used j u st as the product in ord nar algebra It has no peculiar d ffi culties If two vectors A and B are e x pressed in terms of the three unit vecto rs i k as Th e y i . , j B : , _ A B 1 B 1 + A2 B + A3 B i . l l i i + r o s + B i + A1 B j i + A2 B o k o j+ A 3 B 2 2 2 i s ky j+ A1 B j j+ A 2 B o k . j+ A 3 a 3 B i k jk o a n k , means of ( 4) th s reduces to By A r If in ir (7) particular A and B are unit vecto s the components A 1 A 2 A 3 and B 1 B 2 B 3 are the dire cti on cosines of t h e l nes A an B efe e t o X Y Z i , , d r rr d , , , , . , P DIRECT AND SKE W R OD UC TS OF VECTOR S cos (A X) cos (B X) AI B 1 , , , , i cos (A cos (B A2 B 2 , Y) A8 , Y) B , , cos (A Z) c os (B Z) , 3 c os B A , ) ( cos (A X ) cos X) (B , , , . , Moreover A B s the cosine of the included an gle the equation becomes o 59 Henc e . cos ( A Y) cos ( B Y) cos (A Z ) cos ( B Z ) , , , , . case A and B are perpendi cular this reduces to the we l l known relation In 0= cos ( A X ) cos ( B X) , cos ( A Y) cos ( B cos ( A Z ) c o s ( B , , , i be t ween the direct on cosines of th e line A and the line B If A and B are t wo sides O A and O B of a tri angle 0A B the thi r side A B is C B A ( Fig Y) , , Z) . d , FIG 1 7 . . . A— 2 = A + B 0 2 That 2 — 2 A B c os 2AB c os (AB ) . is the square of one side of a t ri angle is equal to the s u m of the squa es of the other t w o sides diminished b y tw c e their product times the cosine of the angle b etween them O r t h e square of one side of a tri angle is equal to the sum of the squares of t h e other two sides di minished by twice the projection of either of those sides upon the other the theorem sometimes known as the generaliz ed P ythagorean theorem If A and B are t wo sides of a parallelogram C A B and D A B are the diagonals Then , r i . , , . , . D = (A — B ) D- o (A — B ) =AA — -A + B 2 AB + B B) . o B, VECTOR A NAL YSIS 60 That l r ll l is the sum of the squares of the di agona s of a para e o gram is equal to t wi ce the sum of the squa es of t wo sides In like mann er als o , D= 4 A CC — D2 — 0 That D = 4 A B 2 . B o c os l is the difference of t h e squares of the di agona s of a parallelogram is equal to four times the product of one of the sides b y the projection of the other upon it If A is any vector ex pressed in terms of i k as , . , j , if A be e xpress ed in terms of any thre e non -c oplan a r un it vecto rs a b c as B ut , , A= + bb + a a a c o + 2 b c b c + 2 2 c a c cos c a o ( This h s c, a ) . y formula is analogous to the one in Cartesian geometr whi ch gives the distance between two points referred to oblique axes If the points he x 1 f’l v z ] a n d x 2 y 2 x2 the d stance squ ared is i 2 D fi , , . 2 ( x2 (3 2 x ) (3 12 1 2 (h (z 2 ( (x2 x z 1) ? — z 1 2 “ ’ I l r) l “ z ) ( x2 S I 19 0 a (z 003 , z 2 1 ) , 2 ( X Y) , Y cos Z ) ( l ) x l ) cos ( Z X) ’ , . of the vector A in to the vecto r B s the vec to r quant t 0 whos e di rection is the norm a upon that side of the p an e of A and B on which De l n i tio n : i Th e s ke w p ro duc t iy l P DIRECT AND SKE W ROD UCTS OF VECTOR S 61 rotation from A to B through an angle of less than one hun dred and eight degrees appears positive or counter clockwi se ; a n d whose magni t ude is obtained b y multiplying the product of the magni tudes of A and B by the s i n e of the angle from A to B a so be de fi ned as that in B ma Th e direction of A which an ordin ar right-han ded screw advances as it tu ns so as C Ax B t o carr A toward B ( Fig Th e skew product is denoted by a cross as the di rect pro uct was b a dot It s written Fm 1 8 y . y y x y r y l ‘ : . d i . . . C= A X E and re a d A c ro s s B Fo r this reason it is often called the c ro s s product More frequently however it is called the ve c to r prod owing to the fact that it is a vector quantity and in con uc t tra st with the direct or scalar product whos e value is scalar Th e vector product is b de fi nit on . , . , , y i . ( 9) r when A and B a e the magnit u des of A a n d B respec t ively and where c is a unit vector in the di rection of C In case A and B are u nit vectors the skew product A x B reduces to the unit vector 0 multi plied by the sine of the angle from A t o B Obviously also if either vector A or B is multipl ied b a scalar x their product is multiplied b that scalar . y y ( If A and B xA ) D= A> < . . ( xB ) = xC . are p a ra lle l the angle between them is either z ero or one hund ed and eighty degrees In either c a se the sine vanishes and consequently the vector pro uct A x B s a nu vecto r An d con versely if A x B s z e ro r . i . AB s in d i ll VE C TOR ANA L YSIS 62 Hence p a ra or B or sin ( A A , lle lis m t wo f o ve c to rs y B) is z ero n e i th e r Thus the . whic h f O c o n di tio n va n is h e s y fo r x is A B a corollar the vector product of an vector into itself vanis hes Th e vector product of two vectors w l appear wh e r 32 ] ever the sin e o f the included an gle is of importance just as the scalar product di d in the case of the cosine Th e two prod They have been u c t s are in a certa i n sense complementary denoted by the two common signs of mul tip cation the dot and the cross In vector an alysis they o ccupy the place held by the trigonome t ric functions of scalar analys is They are at the same time amenable to algebraic treatment as will be seen later A t present a few uses of the vec tor product ma be cited IfA and B ( Fig 1 8 ) are the two adjacent sides of a parall el ogram the vector product 0 . As il . , . . li , . . y , . . . (A B) , c represent s the area of that parallelogram in magnitude and di rection ( Art This geometric representation of A B s of such common occurrence and importance that it might well be taken as the de fi nition of the product From it the tri gonometric de fi ni tion follows at once Th e vector product appears in mechanics in connection with couples If A and A are two forces forming a couple the momen t of the couple is A X B prov ded only that B is a vector drawn from a n y poin t of A to an point of - A Th e product makes its appearance again in considerin g the velocities of the indiv d ua l particles of a body which is rotating wi th an a n u l a r ve g l o c it y given in magn tude and di rection b y A If R be the radius vecto r drawn from any point of the ax is of ro t ation A the product A B will give the velocit o f the e xtremity of R ( Art This velocity is perpendi cular a ke to the a xis of rotat on and to the radi us vector R i x . . . . i y , i . ' i x . i . y . . li KE W PR OD U vector produc s x DIRE C T AN D S T VECTOR S C S OF 63 and B x A are not the same They are in fact the negatives of each other Fo r if rotation from A to B appear positive on one side of the plane of A a n d B rotation from B to A wil l appear positive on the other Hence A B is the normal to the plane of A and B upon that side opposite to the one upon which B A is the normal Th e magnitudes of A B and B A are the s ame Hence Th e A t B . . ' x , . . A Th e f if the Th in a c to rs a f the s ign O x ve c to r p B B ro du c p ro du c t be t x x ca n x x . A ( 1 0) be i n te rc h a n ge d f i a nd o n l y re ve rse d. is is the rst i stance in w the laws of operation in vector analysis differ essentially from those of scalar a n a l y s s It may be that at fi rst this change of sign whi ch mus t accompa ny the interchange of factors in a vector product will ive ri s e to some di f fi culty and co n fusion C hanges similar to g this are however ver familiar No one would think of inter changing t h e order o fx and y in the e xpression sin ( x y ) without p re fi xin g the negative sign to the resul t Thus i fi hi c h n . , y , . . . sin ( y — x) — sin ( x y ) although the sign is not required for the case of the cosine , . cos ( y — x) c o s ( x — y) A gain if the cyclic order of the lette s A B C in the area of a tri angle be changed th e area will b e changed in sign ( Art . r " , . AB C — A CB r the same man ner this reversal of sign which occu s when the order of the factors in a vector product is reversed will appear af t er a little practice and ac uaintance just a s natural and convenient as it is necessar Th e distributive law of multiplicati o n holds in t h e case of vecto r products just as in ordi nar algebra — e xc ep t In , y q . y , VECTOR ANAL YSIS 64 f orde r o th a t the the f mu s t a c t o rs u lly f be c a re m a in t a in e d (1 1 ) 0 A very simple proof ma y be given by making use of the ideas developed in Art 26 Suppose that C is not coplanar with A and B et A and B be two sides of a triangle ta ken in order Then ( A + B ) wi ll be the Form the prism th ird side (Fig of which thi s triangle is the ba se and of which 0 is the slant height or edge Th e areas of the lateral faces of th s pri sm are L . . . . . FIG . 19 , i A x c, Th e . B x c, areas of the bases are x A B ( - A x B — 3 and ( ) ) 2 . the sum of all the faces of the prism is z ero ; for the prism is a closed surface Hence B ut . l 2 (1 1 ) i relation is therefore proved in case 0 s non-coplanar with A and B Shou ld C be coplanar with A and B choose D any vector out of that plane Then C D also w ll lie out of that plane Hence b y ( 1 1 ) Th e i . . , , . Since the thr ee vectors in each set A C D and B C D and A B C D will be non -c oplanar if D is properly chos en the products may be e xpanded , , , , , , , , , . VECTOR A NA L YSIS 66 This may be wri t ten in the form of a determi nant as Ax B i j k AI A, A3 formul ae for the sine and cosine of the sum or dif ference of two angles follow immedi at ely from the dot a nd cros s produc t s et a and b be two unit vectors lying in the plane If x b e the angle that a makes with i and y the ij angle b makes wit h i then Th e L . , . , cos x a i + s in x j ' , — x ) a , a cos (y — x ) b = ’ c os x c os y cos i y s in — s in + y y s in x y s in x. . j , a Hence cos ( y + x ) c os y s in c os x — ) y) — x s in x c o s — Hence sin (y — x) s in Hence sin ( y s in + ) x y c os x — y c os x + s in x c o s y s in x c o s y , . . . and l m n are the direction cosines of two unit vectors a and a referred to X Y Z then If l, m , n ’ ’ ’ , , ’ , cos ( a , , , as has already been shown in A rt 2 9 Th e familiar formula for the square of the sine of the angle between a and a ma be found . . ’ . y RECT DI x a a sin ( a a ) ’ P KE W R OD U CTS AN D S ’ , e = m ( ’ l m ( n - ’ ( ( sin a x a a ) ( 2 (m This a) x n ’ to sin ( a a ) ’ m u ) z + ( l n and a a e , ’ l ’ l m ) k, 2 ’ n 67 ’ where e is a u ni t vector perpen di cular ’ VECTOR S OF sin ( a 2 e ' n ’ l) ’ 2 +(l m , ’ l m) 2 ' leads to an easy way of establishing the useful identity ’ (m n m ’ m 2 n ? u z ( ) ) (l ’2 P ro du c ts ’ n l m ’2 n n ’ l) ’2 2 l m ( ’ l m )2 ’ W ( ) mm ’ n n ’ ) 2 . e h a t r M o t n T w o V ec o rs f O to this point nothing has been said concerning ] products in which the num b er of vecto rs is greater than If three vectors are com b ined in t o a product the result t wo is called a t rip le product N e x t to the simple products A B and A> <B the triple products are the most important All higher products may b e reduced to them Th e simplest triple product is formed b y multiplying the scalar product of two vectors A and B into a thi rd C a s 36 Up . . o . . A ( B)c . This in reality does not di ffer essentially from scalar m ul ti Th e scalar in t hi s case merely happens to pli cat ion (A rt M oreover b e the scalar product of the two vectors A and B inasmuch as t wo vectors cannot stand side by side in the form of a product as B C wit hout either a dot or a cross to uni t e them the parenthesis in ( AB ) C is superfluous Th e e x pression A R0 . . , . 0 cannot be interpreted in any o t her way than as the product of the vector 0 b y the scalar A B 1 . La t e r ( Ch a p V ) g m m p ro du c t B C wh e re n o s i n e i t h e r do t o r c ro s s o c c u rs will b e d e fi n e d B u t it wi l l b e s e e n t h e re t h a t (A B ) C a n d A ( B C) a re id e n t i c a l a n d c o n se qu e n t l y n o a bi u i t y c a n a ris e fro t h e o is s i o n o f t h e pa re n t h e s is 1 . . . th e mg , o , o . VECTOR A NAL YSIS 68 second triple product is the scalar product of two vec t ors of which one is itself a vector product as or X B C A B X C A) ) ( ( Thi s sort of product h as a scalar value and consequently is oft en called t h e scalar triple prod uc t It s properties are perhaps most easily deduced from its commonest geometri cal inte rpretation et A B and C be any three vectors drawn from t h e same origin ( Fig FI G 2 0 Then B x C is t he a re a __o f the par T h e a ll e l o ra m of which B and C are two adjacent sides g Th e , , . . . L , A , . . . __ . A( B x C) 8 03 1“ r ( 1 4) v will therefo e be the volume of t h e parallelopiped of whi ch Se e A rt 2 8 B> <C is the base and A the slant height or edge This volume v is positive if A and D x C lie upon the same side of t h e B C plane ; but negative if they lie on opposite sides In other words if A B 0 form a right-handed or positive system of three vectors the scalar A ( B x C) is posi tive ; but if they form a left-handed or negative system it is negative In case A B and C are coplanar this volume will be neit her posit ive nor negative b u t z ero A n d conversely if the volume is z ero the three edges A B C of t h e parallelo piped must lie in one plane Hence th e n e c e s s a ry a n d s uffi . . . - . , , o , . , , . ‘ , , . c ie n t c o n di tio n f or th e c o p la n a ri ty O f th re e ve c to rs A, B , 0 y n one a corollar the scalar triple product of t hree vecto rs of which two are equal or collinear must van sh ; for any two vecto rs a re coplanar Th e two products AB C and X A B x ) ( ( ) C are equal to the same volume v of the parallelopi p ed whose concurrent edges are A B C Th e sign of the volume is the same in b oth f wh ic h va n is h e s o is i , , . 0 . As . P DIRECT AND SKE W R ODUCTS Of VE C TOR S Thi s equalit may b e state d as a rule of Operation Th e y a nd th e wi tho u t c r o ss a ” . in a l te ri n g the sca la va lu e trip le p r f th e p ro du c t O It m a y als o be seen that the muted cyc licly without altering A( B x C) ro du c t n do t e d g . vecto rs A B the product , may be per C , . ( Cx A ) B m a y be i n t e rc h a 69 1 5 ( ) C each of the e xpressions gi ves the volume of the same pa rallelopiped and that volume will have in each case the same sign b ecause if A is upon the pos itive side of the B C plan e B will b e on the p o sitive side of the C A-plane and 0 upon the pos itive side of the A B —plane Th e triple product may therefore have any one of s ix equivalent forms Fo r , , . A( B x C) x A C ) ( x B C ( ) A B x A B )C ( x C A )B ( r . however the cyclic order of the lette s is changed the product will change sign If . A( B x C) 1 6 ( ) is may be seen from the Th H ence fi gure or from the fact that CxB B xc . l t e re d by i n te rc h a n gi n g th e do t o r the c ro s s o r by p e rm u ti ng c y c li c ly t h e o rde r of th e ve c to rs , bu t i t is re ve rs e d in s ign if t he c y c lic o r der be c h a n ge d : A sc a la r trip le p r o duc t is n ot a y . word is necessar upon the sub ject of parentheses in this triple product Ca n they b e omitted without a m b igu ity ? They can Th e e x pression A . . y Ao C can have onl the one interpretation A i It is im p o s the e xpression ( AB ) x C s meaningless sible to form the skew product of a scalar AB and a vector Fo r . VECTOR A NAL YSIS 70 y i A B xC ma Hence as there is only one way in wh ch C b e interpreted no confusion can arise from omitt ing the parentheses Furthermore owing to the fa ct that there are and C which have the same s ix scalar triple products of A B value and are consequently generally not worth distinguish ing the one from another it is often convenient to use the symbol . , . , , , [A B y to denote an one of the [ A B C] s ix C] equal products AB xc BCx A AxB C A B x C- B A C] [ . C AxB o B CxA- [A C B ] . j scalar triple products of the three unit vectors i k a l l vanish e x cept the two which contain the three different vectors Th e , , . [i = k j ] Hence if three vectors as B = then = A B A B C [ ] I — A1 B A, B a 1 7 ( ) , C i + be ex pressed in terms of j , k a j+ a n , C3 + B , C, B — i, 1 C, A 3 + 01 A , B 3 1 C, A , , — CI A 3 B . Thi s may b e ob tained b y actually performing the mul t iplica tions which are indicated in the triple product Th e result may be wri tten in the form of a determinant 1 . . Al A 2 A3 [A B 1 Th i s is t h e d ro n o n e e x e rc i se s t e t ra h e see . m l gi i fo r of u a wh o s e ve n C] B , B B , , 01 02 03 in ve rt c e s is so lid at the g m y g m g ly t i c o ri i n a na eo . Fo r e tr a o re mm l l e e n e ra l f o r fo r t h e vo u o fa u a DIRECT KE W PR OD UCT AN D S S OF Ifmore generally A, B non-coplanar vectors vectors , where a s ta n t s , then l , a ,, a s C , b, 3, A= a C= C ; I I a + a a + C 1 ba c , b1 — 0 “ [A B b + a b + C 2 2 and b l , b, , b, ; c a a , 3 I c + 3 l c c y 3 0 “ — c a , l 3 “ “ 2 3 o 2 c ,, a 0 1 1 0 3 c , b 1 b, b8 C] 71 are e x pressed in terms of any three c which are not necessarily unit [ A B C] — VECTOR S [a 3 a re certain con bs a b ,) b c [ a b c ] . ] s third t pe of tri ple product is the vector product vectors of which one is itself a vector product Such Th e . and vector is pe rpendi cular to A and to (B x C) B u t ( B x C) is perpendicular to the plane of B and 0 Hence being perpendicular to ( Dx C) must lie in the plane of B and C and thus take the form Th e . . s a y c, where x and y are two scalars In like manner also the vec tor being pe rpend cular to must li e in the plane of A and B Hence it will b e of the form . i . =mA where m and v are two scalars in general is . From n o t e qu a n B i this it s evident that l to parentheses therefore cannot be removed or inter changed It is essential to know which cross product is Th e . VECTOR A NA L YSIS 72 formed fi rst and which second This product is termed the vector triple product in contrast to the scalar triple product Th e vector triple product may be used to e xpress that c o m n t of a vector B which is perpendicular to a given vector n e o p This geome t ric use of the product is valuable not only in A itself but for the light it sheds upon the properties of the product B AX et A ( Fig 2 1 ) be a give n ve c t o r and B another vector whose com n t s parallel and perpendic u lar n o e p A to A are to be found et the components of B parallel and per n dic u l a r to A be B e and r e B p Draw A and B from a s p e c t ive l y common origin Th e product AxB Th e product is perpendi cular to the plane of A and B It is furthermore Ax ( AxB ) lies in the plane of A and B perpendi cular to A Hence it is collinear with B An e xamination of the fi gure will show that the direct on of Hence AX ( AxB ) is opposite to that of B . . . L . ~ . . L ” ’ . . . . ” i . . ” . c where c is some scalar constant Now but ‘ i B ” — c , . sin (A B ) b = 34 B — c B ” 2 , in (A , B) b if b be a un t vector in the direction of B ” Hence c Hence A2 ” ” , ” A A- . AA component of B perpendicular to A has be en e x pressed in terms of the vector triple product of A A and B Th e component B parallel to A was found in Art 2 8 to be Th e , , ’ . . VE C TOR ANAL YSIS 74 Substituting these values in B C A- AC B A> <(Ex C) ( 24 ) . relation is therefore proved for any three vectors A B 0 A nother method of giving the demonstration is as foll ows It wa s sho wn that the vector triple produc t Ax ( B XC) was of t h e form Th e , . , . . Ax ( B XC) Since Ax it b y A is z ero . B C y is perpendi cular to Hence A, t h e x A and Hence o Ax ( B x C) where n is a scalar constant Mult iply by B . 0 AB . A ( CB n direct produc t of A C y B AC w y . AB C), remains to show It 1 71. . . n (A c B B AB 0 o o scalar triple product allows an interchange of dot cross Hence Th e and . if the order A( B x C)xB of the AB C B [ — B C Ao B o AC B and 1 n A factors ( B x C) and B be inverte d A[B x ( Dx C) ] Hence — From t h e y BB . 0] BB AC o . J ADC . ( 24 ) three letters A B C b difi e re n t arrangements four allied products in each of whi ch B and C a re included in parentheses may b e formed These are , , , . a vec t or product changes its sign whenever the order of two factors is intercha nged the above products evidently satisfy the equations As , A> <( B ) xC D IRECT KE W AN D S P S R OD U C T VECTOR OF S 75 e xpan sion for a vector triple product in which the paren t hesis comes fi rst may therefore be o b tained directly from t hat already found when the parenthesis comes last Th e . A B C- B A C- Th e . form u l ae then be come = AC B AB C AC B 3 A 0- and These reduction formul ae are of such constant occurrence and great import ance that the should b e com m i t ted to memory Their content may be stated in the followi ng rule To e xp a n d y fi . . mu l t ip ly th e e xte rio r fa c to r i n t o th e p re m o te r t e rm in t h e p a r e n th e s is to fo rm a s c a l a r c o efii c ie n t fo r the n e a re r on e th e n m u l t ip ly th e e xte ri o r fa c tor i n to th e n e a re r a vec tor t rip le r o du c t rs t , t he t e rm i n pa re n th e s is re mo te r o n e , a n d s u btr a c t to th is fo rm a res u l t sc a la f r th e ro m fi c ie n t fi c oe rs t f or t he . far as t h e prac t ical applicati ons of vector analysis are concern ed one can generall get along without any formul ae more complicate d than that for the vector t riple product B u t it is frequently more convenient to have at hand other reduction formul ae of which all may b e derived simply b y making use of the e xpansion for the t riple product and of the rul es of operation with t h e triple pro duc t AB xC To reduce a scalar product of two vectors each of which is itself a vector product of two vectors as As y , . . , (AxB ) L et this be regarded as a scalar t vecto rs and thus ri A, B , <D C> - AxB 4 Cx D) Inte rc ha nge the dot and the cross . r ple product of the th ee VE C TOR ANAL YSIS 76 AxB ( CxD) B x ( CxD) A- BC D B B 0 Hence This y in D B C A- D AC B- A B x ) ( Cx D) ( ma be written . . determin antal form 2 ( 5) . AD D A B O X x )( ( 3 4) and D b e call ed the e xtremes B and C the means A and C the antecedents ; B and D the consequents in this product according to t h e familiar usage in proportions then the e xpansion ma be stated in words Th e scalar product of two vector products is equal to the ( scalar) p fo du c t of the antecedents times the ( scalar ) product of the conse uents di minished b y the ( scalar) product of the means times the scalar product of the e x tremes ) ( To reduce a vector product of two vecto s each of whi ch is i tself a vector product of two vecto s as If A y , . q r . r Let C> <D Th e E . product becomes AE B Substituting A ( CxD) B F AxB B EA . the value of E back into the equati on o L et , Th e . ( 2 6) ( B Cx D) A o produ ct then becomes ED 0 A B x D)C ( q FC D A B x C) D ( . equating these two e ui valent results and transposing all the terms to one side of the e uation By [B C D] A q , 2 ( 7) Th is is an equation with scalar coe ffi cients between the four vectors A B C D There is in general o nl y one such equa , , , . P DIRECT AND SKE W R OD UCTS OF VECTOR S 77 y tion because any one of the vectors can be e xpressed in onl one way in terms of t h e o t her three : thus the scalar c o e ffi c ie n t s of that equation which e xists between four vectors are found to be nothing b u t the four scalar tri ple products of th o s e vectors ta ken three at a time Th e equation ma a so b e written in the form , yl . c M ore . e x amples of reduction formul a of whi ch some are import ant are given among t h e e xercises at the end of the chapter In view of these it becomes fairly obvious th at the com binat ion of an numbe r of vecto rs conn ected in any legitimate wa b y dots and crosses or the prod uct of an numb er of such combinations can b e ultimately reduced to a sum of terms each of which conta in s onl y one cross at most Th e proof of this theorem depends solely upon analyz ing the possible combinations of vectors and sho wing that they all fall under the reduction formul a in such a way that t h e crosses m a y be removed two at a time until not more than one remains Th e formul a developed in the foregoing article have interesting geometri c interpreta tions They also a fi o rd a simple means of deducing the formul a of Spherical Trigo n o me tr T hese do not occur in the vect o r a n al sis proper y The ir place s taken by the two quadruple products , , y . y y . . . . y i . , D A0 B- ( AxB ) i L B e An [A CD] B [B [AB D] c A B C] [ ( 2 5) e n] A 2 ( 6) wh ch are now to be inte rpr eted et a unit sphere ( Fig 2 2) be g ven et th e vectors A B C D be unit vectors dr awn from a common origin the centre of t h e Sphere and terminating in the surface of the Sphere at the points A B Th e great circular arcs C D , i . . , , . L , , , , , . VECTOR ANAL YSIS 78 AB A A , . and C , r etc give the angles between the vecto s A a n d B etc Th e points A B C D determine a quadrilateral upon the sphere A C and B D are one pair of opposite sides ; A D and B C the other A B and CD are the diagonals , , , . , , . , . . AC B D AB B C o sin ( A B ) CxD sin ( 0 D) Th e angle b e t ween A> <B and CXD is the angle b e t ween the normals to the AB Fu r 22 and CD planes This is the same as the angle between the planes thems elves et it be denoted by x Then AxB , , . , . . - L . . . sin (A B ) sin ( C D) cos x A B x ( ) angles ( A circular arcs A B Th e B , , y . , , ma b e replaced by the great which measure them Then C ( ) , , CD D) . sin A B sin C D cos cc cos A C cos B D cos A D cos B , AC B D — AD BC o O . Hence s in AB s in CD c o s cc c os A C o os B D — cos AD cos B C . words : Th e product of the cosines of t wo opposite sides of a spherical quadr ilateral less the product of the cosin es of the o t her two Opposite sides is equal to the product of t h e sines of t h e diagonals m ul tiplied b y the cosine of the angle b etween them This theorem is credited to G auss et A B C ( Fig 2 3 ) b e a spherical tri an gle the sides o f which are arcs of great by a b c circles et the sides be denoted F G 23 respectively et A B C b e the unit vectors drawn from the center of the sphere to the points A B 0 Furt hermore let p p p be the great circular arcs dr opped In . L . , L , . I . . , . . L , , , , , ,, , , , , . P DIRECT A ND SKE W R OD UCTS OF VECTOR S perpendicularly from the vertices A Interpret t h e formula B , C , sin ( A B ) s in c , , Then a , b, c . BA C A- AC B A o to the sides 79 o . sin (0 A) sin 0 sin b cos x A x C ) ( sin , b . , where is the angle b et ween A><B and C><A This an gle is equal to the angle b etween the plane of A B and t h e plane of C A It is ho wever not t h e interior angle A which is one of the angles of the triangle b u t it is t h e e x terior angle 1 8 0 A as an e x amination of the fi gure w l show Hence sin 0 sin b cos ( 1 80 A ) sin c sin b cos A Bcos b cos c cos a 1 AA C BC AA B y equating the results and transposing cos a c o s b c o s c — s in b s in c c o s A cos b cos c cos a sin e sin a cos B cos c cos a cos b s rn a s i n b cos 0 Th e last t wo may b e obtained by cyclic permutation of t h e letters or from t h e iden t i t ies x . , , , . , il : ° , . ° . , . BA CB A, C- B AC C- BC o y N ex t . inte rpret the identit in the special cases in which one of the vecto rs is repeated . L et the three vectors a B x C, C> <A, Ax B B A A C [ ] . be unit vectors in the respec t ively Then , c di rec t ion of . Ax B = c s in c , — Ax C b s in b sin 0 sin b A sin 0 sin b sin A C c C sin 0 cos ( 9 0 -p ) sin c c [ A B C] b, xb ° ~ = A B A Sin C [ ] c c s in p; 80 By VECTOR A NAL YSIS q e uatin g the results and cancelling the common factor , siu p sin p sin p , sin b sin A sin 0 sin B sin a sin e a y last two may b e obtained b y c clic permutation le t t ers Th e formul a give t h e sines of the altitudes of the tri angle in terms of the sines of the angle and sides wri te Th e . . A B A C [ ] B A B C [ ] [0A B ] Hence sin c sin b sin A sin a s in c s in B sin b sin a sin 0 . A B C [ ] . i 0 : B [ C A] : [C A B ] . q l express ons [ A B C] [ B CA] [ C AB ] are e ua the results in pairs and the formul a Th e , s in , b s in A = s in s m c s rn sin a sin s Cz a s in . B in b s in C s in c s in A vectors is z ero for the triangle is a clos ed polygon From this equation as . VECTOR ANAL YSIS 82 where a b M ultiply b y , are thre e scalar constants to be determined c , . b xc . t) b xc r a a [ In a o r b bb xc b xc h o ] [a a b c c c ] b xc . like manner b y multiplying the equation by b the coeffi cients b and c may be found x c . = b b [ [ ] [r a b ] re a = [r b c ] a a b c [ ] Hence [ c c a [ b c a] denominators are all equal equation Th e [a b c ] b] c a . c a [ b] Hence this c ]r which must e xist b etween the four vectors r a Th e equation may also be written , r ~ b > < [a r . c c ] b x c b ' r c > < a [ b a c ] c > < [ b b r o [a a b c ] a c ] b, x b a b c c ] a x b [ b . [a , a c ] c . x a and DIRECT A ND SKE PV PR OD UCT OF VECTOR S S 83 which are found b y di vidi ng the three vector products b x c c op la n a r vectors a b c b c x a a x b of t hree n o n y the scalar product [ a b c ] is called t h e re c ip ro c a l s y s te m to a b c If a b 0 were c o Th e word n o n -c op la n a r is impor t ant plan ar the scalar triple product [ a b 0] would va ni sh and consequently the frac t ions c x a b x c a b , , , , , , . , . , x [ a b c ] [ b a [ ] c b a c ] would all b ecome meaningless Thr ee coplanar vectors have no reciprocal system This mus t b e carefully rememb ered Hereafte r When the term re c ip ro c a l s y s te m is used it will b e understood that the t hree ve c to rs a b c are not coplanar Th e system of three vectors reciprocal to s stem a b 0 wil l be denote d b y primes as a b c b x c a x b c x a . . . , , ’ ’ [ Th e e xp re s s io n b ] c fo r [ b a c . , , ' , , a y , [ ] a b c ] r reduces then to the very simple form ’ r a a ’ r c c 3 r r b b 0 ( ) The vector r may be e x pressed in terms of the re ciprocal system a b 0 ins t ead of in terms of a b c In the fi rst place it is necessary to note that if a b c are non-coplanar a b c which are the norm als to the planes of b and c c and a a and b must also b e non-coplanar Hence r may be e xpressed in terms of them by means of proper scalar coefficients a y z ’ ’ , ’ o ’ , , ’ , , . , , , ’ . ’ , , , . , , . b or c a r [ ] Multipl successively b y b a [ 0] N i y [ ] [a b c ] a H ence b r c o a , o x b, o b [ c c This . a] , ’ r a a z [a b c ] r o b b ’ - 33 r ro c gives , r a. = y ro b 2 re c ’ r c c , . VECTOR ANAL YSIS 84 a b c the s stem reciprocal to the y ] scalar product of any vector of the reciprocal system in to the correspon ding vec t or of t h e given system is unity ; b u t the product of two n o n -corresponding vectors is z ero ’ If a 44 b , ’ ’ c , be , , . ’o a a ’ o b ’ a o ’ a b b c ’o ’o b c ’ o b a 1 c o c c ’ c a o ’o 0 b . This may b e seen most easily by e xpressing a terms of themselves according to the formul a ( 3 1 ) r Hence a b o Since ro a a ’ ’ a ’ b ’ ’ c a aa o o ro b ’ a o ’ ’ a ’ c b b bb ’ b b ' o ’ o b , c ’ in . ’ o ’ b ’ a a ’ , ’ ’ r c c ’ b ' ’ a c c c o ’ b ’ ' o c c c c ’ ’ ’ . are non-coplanar the corresponding c o e ffi c ie n ts on the two sides of each of these three equations mus t b e equal Hence from the fi rst a ’ , b ’ c , ’ . From the second Fro m the t hird ’o 1 a 0 Na 0 ’o c O a This a 1 ’o ’ c o O b o b 0 a ’ a O b c 0 . b 1 b ’ ’ o ’o c c . proves the relati ons They may also directly from the de fi n i tions of a b c . , b > < c [ ] b a c b > < c [ ] a b c -a a b c ] b x c o h [ a proved , c [ be ’ ’ ’ . b c b [ o a [ b a ] c] 0 ] [a b c ] and so fort h Conversely if two sets of three vectors each say A and a b c satisfy the relations . , , , , Aa o A 0 3 A 0 0 =B h o B o a Cc “ 2 B 0 0 z 1 03 0 Cb = O o , B , C, VEC TOR A NA L YSIS 86 Th eo re m ’ Ifa b , ’ c , scalar triple products [ a reciprocals That is ’ the and a ’ ’ b ] y reciprocal s stems c ] are numeric a l b a [ be c an d ’ c b, , . [ a [ c b x [ ’ a <a c> b b ’ ] [ ’ c b x [ b a ] c 3 a b c c c ] [b x c 1 ] 0 x c > <a [ b ] [ c xa a x b ] b c a c ] . xb ] a [ a b c ] c [ xb ] a x b a b a c b x [ a a 1 Hence b [ a c 3 ] [ a b o c c ]c [ xb a 2 ] . b a c ] 2 . 1 _ [ b a c ] means of t his relation b etween [a b c ] and [ a is possible to prove an i m portant reduc t ion formul a ’ By ’ ’ b c ] it , A P- P B PC A Q- Q B C Q- BA B B B C o o whi ch replaces the two scalar triple produc t s by a sum of nine terms each of which is the product of three direct pro ducts Thus the t wo crosses which occur in the two scalar products are removed To give the proof l e t P R be e xpressed as . . P , =P A A ~ B P- ’ B A A + QB B Q = Q’ = B B A A B B B ’ . [P e n ] : o o ' P C C o ’ 610 ’ PA P B PC A Q- a s we RA B B - RC I [A B I o C J[A M T I 1 ’ 0 ' c ' . Q , DIRECT Hence f s y s te m o [ k B P- A Q- Q B B A B B i u n it r e c t o r s , j k is , r e c ip ro c a i 1 xj k [1 1 k] Fo r thi s 87 o i k 1 ] 1 t hre e VE C TOR S OF PA [ A B C] [P Q R ] jx P KE W R OD UCTS AN D S l 35 . reas on the primes i j k are not needed to denote k Th e primes will a syste m of vectors reciprocal to i therefore b e us ed in the fut ure to denote anot her set of rect angular ax es i k jus t as X Y Z are used to denote a set of a xes difi e re n t from X Y Z ’ ’ j , , , j ’ a re onl y ri h t -h a g t he Let fth re e sy s te m s o ve c t o rs whi c h n de d l eft -h a a nd j , n de d o wn re c i p s y s te ms f O , 00 B B o 1 Hence the vectors are all unit vectors AB o A t h e ir , 32 ( ) AA Hence a re ro c a ls th r ee j u n it is the system i k and the system i — k C b e a set of vecto rs which is its own reciprocal A, B , Then b y . . That r e c to rs . ’ , , , Th e ’ , , . , is perpendicular to BA Hence B is perpendicular to A C- e and BC A a nd . . . 0 Ac B . , . C . 0 . C . B C- 0 . Hence C is perpendicul ar to A and B k Hence A B C must be a system like i k or like i A scalar equation of the fi rst degree in a vector r is an equa t ion in each term of which r occurs not more than once Th e value of each term must be scalar As an e xam ple of such an equation the following may be given . , , , j , , . j . . a a b xr c fr o cl O, , . VECTOR ANAL YSIS 88 where a b c d e f are known vec t ors ; and a b c d known scalars O bviously an scalar e q uation of the fi rst degre e in an unknown vector r ma b e reduced to the form , , , , y , . , , , , y A= a r where A is a known vector ; and a a known scalar To a c compl ish th is result in the case of the given equation p rocee a s follows . a a { xb a a r b fr c b xb d . , r f} - c 0 d o d . more complicate d forms it may be ne c essary to make use of various reduction formul a before the equati on can be mad e to take the desired form In . , A a r . a vector has th ree degrees of freedom it is clear that one scalar equation is in s u fii c ie n t to determine a vector Three scalar equa t ions are necessary Th e geometric in terpretation of the equa ti on As . . A r ( 36 ) a L is interesting et r be a variable vecto r et ( Fig 24 ) drawn from a fi x ed origin A b e a fi xed vector drawn fr om the same e uation then becomes . . FIG 24 . origin . . Th e . q 7 A r cos ( r A) ‘ COS a 3 , , , if r be the magnitude of r ; and A that of A . r is L cos ( r , Th e r e xp ession A) the projection of r upon A Th e equation therefore states that the projection of r u pon a cer t ain fi x ed vector A must . VECTOR ANAL YSIS 90 From q four scalar e uations ro A = ro B = b C= r I 'D a c =d li the vector r may be entirely eliminated To accomp sh this solve three of t h e e q u a t io n s and su b s t itute the value in the fourt h r . ’ . a a ' A D o B D O [ ] bB ’ o D [ C A D] b c ’o d C Dz = D d A B ] [ c [ A B C] ( 4 0) . vector equati on of the fi rst degree in an unknown vector is an equat ion each term of which is a vector quantity con t aining the unk n own vector not more than once Such an equation is A . D Er n r F where A, B , C, D, E, F 0, are known vectors n a known scalar and r the u nknown vector O n e such equation may in gen eral b e solved for r That is to say one vector equation is in general suffi cient t o determine the unknown vector which is con ta ined in it t o the fi rst degree Th e me t hod of solving a vector equation is to multiply it wi th a dot successively b y three arb itrary kno wn non-00planar vectors Thus three scalar equations are Obtained These may be sol ved b y the methods of the foregoing article In the fi rst place let the equation b e , , . . , . . . . Aa where o r r + C c + B b- ~ r - D, a b c are known vectors No scalar c o e ffi c ie n ts are wri t t en in the ter m s for th ey may be incorporated in t h e vectors Multiply the equation successively b y A B C It is understood of course that A B C are non00planar A, B , C, D, , , . , . , , , , ’ ’ ’ . DIRECT A ND PR ODU SKE IV r b- ’ DA ’ DB c r DC a r o o B ut a ar ’ ’ DA a ’ r Hence r T OF VECTOR S 91 C S ' o r bB b D- ’ b ’ ’ c c or . DC ’ ’ c ’ . solution is therefore accomplished in case A B C are non oplanar and a also non coplanar Th e special cases in b 0 C which either of t hese sets of three vectors is coplanar will not b e discussed here Th e mos t general vector equation of the fi st degree in an unknown vec t or r contains terms of the t pes Th e , , , , . r . y A - I) E xr, n r, a r, . That is it will contain terms which consist of a known vector multiplied b y t h e scalar product of ano t her known vec tor and the unknown vector ; t erms which are scalar multi ples of t h e unknown vector ; terms which are t h e vector product of a known and the unknown vector ; and constan t r may al ways b e reduced Th e terms of the type A a t erms to thr ee in numb er Fo r the vectors a b c -which are multiplied into r may all be e x pressed in terms of t hree non coplanar vec t ors Hence all the products a r br c r may be ex pressed in terms of three Th e sum of all terms o f r therefore reduces to an e x pression of thr ee t h e t ype A a terms as . ’ , . , , , . , o , . , r Aa- y terms of the t pes in thi s form Th e r B b- r and n Cc E> <r o r . may also be e xpressed . n r E xr A dding n a ’ Ex a a r o ’ + a r o n b ’ r + b’ Exb b n c ’ c or c or q . all these terms together the whole e uation reduces to the form r KL o ’ a ’ ’ Km b ’ K r solution is in terms Oft hree non coplanar vecto s a b c These form the system reciprocal t o a b c in terms of which t h e prod u cts con t ai n ing the unknown vector r were e x pressed - Th e , ' , ’ , ’ . , . SUN DRY APPLICA TIO N S OF P R OD U CTS Ap p lic a tio n s to Me c h a n ic s the mechanics of a ri gid b o dy a force is n o t a vector in the sense understood in this book Se e Art 3 A force has magnitude and direction ; b u t it has also a line of application Two forces whi ch are alike in magnitude and direction but which lie upon different lines in the bod do not produce the same e fi e c t N evertheless vectors are su ffi ciently like forces to be useful in treating them -a c t on a body a t the If a number of forces f 1 £ 2 £ 3 same point 0 the sum of the forces adde d as vecto s is called the re su lta n t R 48 ] In . . . y . , . . , , r , , . the same way if f 1 £ 2 f s do n o t act at the same point t h e term resul t ant is still applied to t h e sum of these forces added just as if they were vectors In , , . ( 41 ) a force does not differ from a vector De n i tio n : Th e m o me n t of a force f about the poi n t 0 is equal to t h e product of the force b y the perpendic u lar dis tance from O to the line of action of t h e force Th e moment however is b est looked upon as a vector quantity Its m a g n it u de is as de fi n ed above Its direction is usually taken to fi . . . . VECTOR ANALYSIS 94 f T h e direction M o his is the magnitude of the moment T {} of dx f is t h e same as the direc t ion of the momen t Hence t h e relation is proved d M O {f} . . . . sum of t h e moments ab out 0 of a number of forces ac t ing at t h e s a m e p o in t P is equal t o the moment f1 f2 of t h e resul t ant B of the forces ac t ing at that point Fo r let Then d b e t h e vector from O to P Th e , , . . total moment ab out 0 of any numb er of forces f 1 f2 ac t ing on a rigid bo dy is equal to the to tal momen t of those forc es ab out 0 increased b y the moment ab out 0 of the result ant B o considered as acting at 0 Th e ’ , , ’ . M o {fl ’ £ 2 Let M O {fv f2 9 , M O {3 ' 0 4 4 ( ) 2 be vectors dra wn from O to any point in f1 f2 respectively et (I d z b e the vectors drawn fro m O to the same points in f 1 £ 2 respecti vely et 0 b e t h e vector from O to O Then d v dz , , L . , ’ , ’ , . , L ’ . Mo {f1 , f2 , d1 > <f1 M O {f1 , f2 , (l ' d z x f2 ’ f x l l (d1 ( d2 n B is the vec or drawn from O to 0 Hence — C x f is the moment ab out 0 of a force equal in magnitude and parallel in direction to f1 b u t situate d at 0 Hence ut c ’ t . ’ . , DIRECT AND SKE W H ence PR OD UCT S c M 01 {f1 , f2 , MO Ma o f £ { p 2 OF VE C TOR S r {R0 } Mo , i 95 . R { o ( 44 ) i theorem is therefore proved Th e resul tant R is Of course the same at a ll points Th e subs cript O is a t t ached merel t o sho w at wha t poin t it is s up p o s e d to act when the moment about 0 is taken Fo r the point of applicat ion of B affects the value of that moment Th e scalar product of the total moment and the resultant is the same no matter ab out what point the moment be taken In other words the product of the total moment the result ant and the cosine of the angle b etween them is invariant for all points of space Th e . y . ’ . . . , , . £2 , where O and O are any two points in space This important relation foll ows immediately from the equation ’ . MO Fo r R Mo ’ f { l a r § f1 , f2 f2 MO Z 9 ' M o {f1 =B , M O {R 0 } , + R f2 , the moment of B is perpendicular to the point 0 of application b e Hence B ut . B . -M O {B 0 } , . no matt er what . R Mo o {R 0 } ' 0 and the relation is proved Th e variation in the total moment due to a variation of the point ab out which the moment is tak en is always perpendi cular t o the resulta nt A poin t 0 may b e found such that the t otal moment ab out it is parall el to the resultant Th e condi tion for parallelism is . . ’ . x E t ra B XM 0 , if“ £ 2 , ' f { 1 , £2 , = R> <M o 0 0 ( fl + R X MO ' , £2 , zRo = g O VE C TOR ANA L YSI S 96 where O is any point chos en at random Replace f t its value and for revity omit to wri e the f b 1 2 by Then b races . , B xM o B B in , 0 } the . problem is to solve thi s equa t ion for Th e / 0 B xM o B xM o f BI O {B B 0 c . c R 0 . is a known quantity M O is also supp osed to b e kno wn et c b e chosen in the plane through 0 p e rp e n O and the e uation reduces to dicular to R Then RC Ne w B L . . q . RXM O RB O 0 B XM O R R- l chosen equal to th is vector the tota moment abo ut the point O which is at a vector distance from 0 equal to c M oreover since the scalar product of will b e parallel to R the total moment and t h e resultan t is constant and since the res ul t ant i t self is cons tant it is clear that in the case where they are parallel the numerical value of the total moment will b e a mini mum Th e to ta l moment is unchanged by di splacing the point ab ou t which it is taken in the direction of the res ul tan t IfC b e ’ , , . , . . M O § f1 , f2 , FOI ‘ M 0 {f1 , £ 2 , ' OXR o is parallel to B c vanishes and the moment ab out 0 is equal to that about 0 Hence it is possible to fi n d n o t merely o n e poi nt 0 a b out which the total moment is parallel to the resulta nt ; but the total moment about any poin t in t h e line drawn through 0 parallel to R is parallel to B Furtherm ore t h e solution found in equation for C is t h e o n ly one which e xists in the plane perpendicular t o B unl ess t h e result ant B vanishes Th e results that have been ob tained may b e summed up as follows : Ifc ’ 0O , ’ . ’ ’ . . . . VECTOR A N AL YSIS 98 y il q the ax is Th e velocit of a n y po int in its c rc e is e ual to the product of the angular velocit and the ra d us of the circle It is therefo e e ual to the product of the angular velocit a n d the perpend c ul ar d s tance from the point to the axis s Th e direction of the velocit x is and to a rp e n dic u Ia r to t h e e p the radi u s of the circ e described b the point et a ( Fig 2 5) be a vector dra wn along the axis o f rotation in that direct on in which a right-handed screw woul d a d vance f turned in the direction in wh i ch the bo y is F o 25 rotating et the magnitude of a be a the angular velocity Th e vecto r a m a y be taken to represent the rotation of t h e body et r be a rad us vector drawn from any point of the ax is of rotation to a point in th e Th e vector product b ody . y r q . y i i i yi l y L . . i r i L L . . . . , . . d i . ( a r s in a > <r ) a, r y is equal in magnitude and direction to the velocit v of t h e terminus of r Fo r its directi on is perpen di cular to a and r and its magnitude is the product of a and the perpendicular distance r sin ( a r) from the point to the line a That is . , . v a > <r . y If the ( 4 5) body be rotating simul taneous l about several ax es a 3 a which pass through the same point as in t h e ] 2 3 cas e of the g roscope the velocities due to the var ou s o t ati ons are v a > <r , r , y i , 1 l 1 v2 v8 a r x s 3 DIRECT A ND KE W PR OD UCT S r S OF VECTOR S 99 where r 1 r 2 t a a e the radi i vectores drawn from points on the ax is a l to the same point of the body Le t ea the vecto rs r l r 2 t a be drawn from the common point of intersect on of th e ax es Then , , , , i , , . , , . =r ' o o y This shows that the bo d moves as if rotating with the angular velocity which is the vector sum of the angular velociti es a l 8 2 a 3 Th is theorem is sometimes known a s the parallelogram law of an g ul ar velocities It will be shown later ( Art ) 6 0 that the motion of any rigid body one point of which is fi xed is at each ins tant of time a rotation ab out some axis drawn through that point Th s ax is is called the instantaneous a x is of rotation Th e ax is is not the same for all time but constantly changes its pos ition Th e motion of a rigid body one point of whi ch is fi x ed is th ere fore epresente d by , , , . i . . , r . v a 4 5 ( ) xr where a is the instantaneous angular velocity ; and r the radi u s vector drawn from the fi xed point to any point of the body Th e most general motion of a rigid body no point of which is fi x ed may b e treated as follows Choose an arb itrar poin t 0 At any instant this point will have a velocit v Re l a ti ve t o th e p o in t 0 the b ody will have a motion of rota tion ab out some axis drawn through 0 Hence the velocit v of any point of the body may b e represente d b y the s um of of O and e xr t h e velocity of that po int V the velocit relative to 0 v v a xr ( 46) , . y . . y . y 0 . 0 . y 0 . VECTOR A NALYSI S 1 00 case V is p arallel to a t h e b ody moves around a and along a simultaneou sly This is precisely t h e motion of a scre w advancing along a In case V is perpendicular to a it is possib le to fi n d a point given b y the vector r such that its veloci t y is z ero That is In , 0 . , 0 . , , . Thi s a y xr ma be done as follows M ul tipl . y by xa . vo x a - Le t vo x a At a a a r r be chosen perpendicular to a Then . vo aoa r x xa v0 . -is z ero and a r a - a a y point r thus determined has the propert that its ve l o c ity is z ero If a line b e drawn through t hi s point parallel to a the mot ion of the b ody is one of instan taneous rotati on ab out t his n e w ax is In case v is neither parallel nor perpendicular to a it m a y b e resolved into two components Th e , , . , . 0 V0 V0 ' V0 i whi ch are respectively parallel and perpend cular to a v v v a ><r A point may now be found such that v a xr et the different points of the b ody referred to thi s point be denot ed b y r Then the equat ion b ecomes ’ ” o 0 ” L . , ’ . v vo ’ a Th e ’ xr . motion here e xpressed consists of rotation about an a xis a and t ranslation along that a x is It is therefore seen that t h e most general motion of a rigid b ody is at any instant . 1 02 As VECTOR ANAL YSIS l f is paral el to a the scalar product [ a ’ dxi a- a o i van shes df . ” dxf . the other hand the work done by f is equal to the work done by f during the displacement Fo r f being parallel to If h be the common a is perpendic ul ar to its line of action vecto r perpendicular from the line a to the force f the work done by f during a rotation of angular veloci t a for time t is appro ximatel On ’ . . y y W : " hf t a y A " nxf t . y vector (1 drawn from an point of a to any p oint of f ma be broken up into three components of which one is h another is parallel to a and the third is parallel to f In the scalar triple product [ a d f o nl y that component of d which s perpendic ular ali ke to a and f has an e fi e c t Hence Th e , i , y ” W a h xf : . ’ ’ a dxf t t a If a dx ft . rigid body upon whi ch the forces f 1 f2 act be dis placed b y an angular velocity 9 for an infi ni tesimal time t and if d l d2 be the vectors drawn from any point 0 of a to any points of f f respectively then the work done 1 2 by the forces fl f2 will be a pp ro xim Me ly W= ( a d1 > <f1 a d > t 2 <f2 , , . , , , , , , , o a a i q If the d ( l o (l 1 M o {fl , £ 2 , f x g 2 t t . body be n e u ilibrium thi s work must be z ero Hence aM o if] £2 t 0 Th e scalar product of the angular ve oc t 9 and the tota l moment of t h e forces fl f, about an point 0 must be z ero A s a may be any vector whatsoever the moment itse f must vanish . , , , , o : . l iy y . . M o { f1 , f2 , O . l DIRE C T KE W PR OD U AN D S y T OF VE C TOR S C S 1 03 y necessar conditions that a rigid bod he in e q u ili b riu m under the action of a system of for ces is that the resul t ant o f t hose forces and the total moment about any point in spa ce sha ll vani sh Conve sely if the resultant of a s s tem of forces and the moment of those forces about any one particular point in s ace vanish simultaneously the bod will be in e q uilib rium 0 then for an di splacement of tra ns lation D If R Th e r y . , , p y y . DR=O o . W= I) f1 f, + D- y r and the tota l work done is z ero when the bod suffe s any d s placement of tran slation be z ero for a g ven point 0 Then for et M O { i n £ 2 an other point 0 i , L y , ’ ut . ’ M O { i n fa B i . y , by h p othes is Henc ‘ R a e y o MO h is also MO z ero f £ { p 2, : MO £29 . ’ Hence { B O} 0 4t where a is an vector whats oever B u t this e xpression is equal to the work done by the forces when the body is rotated for a time t with an angular velocit a about the line a passing through the po in t O Th s work is z ero An y displacement of a rigid bod may be regarded as a t ranslation through a d stance D combined with a rotation for a time t wi t h angu lar velocity a about a s u itable line a in space It has been proved that the total work done by the forces during thi s di splacement is z ero Hence the forces mus t be in e uilibrium Th e theo e m s proved . ’ . i i y y . . q r i . . . VECTOR ANAL YSIS 1 04 Ap p lic a t io n s t o G e o m e try Relation s b e t ween two right-handed systems of th ree Le t i mutually perpen di cular uni t vectors k and i k b e t wo such syste ms They form t heir own reciprocal systems Hence , . j ’ , , j ’ , . . scalars a l a 2 “3 5 bl direc t ion cosines of i ; ; Tha t is Th e , j , ’ a 1 b1 c cos cos (j i ) cos ( k i) same manner (i ’ , i) In t h e ijj + ’ k jj ’ ’ ' , jj u k kl l 0 , 0 2 3 , , ’ , ’ , ’ ’ i k k j o ’ k k ’ ’ ’ k k k i i ' k ’ 3 c ; , o o l are respectively the wi t h respec t to i k bz , b j j j cos (i ) cos (j ) cos ( k ) 2 0 2 , + ’ b2 , ’ 1 , ’ a ’ ’ 1 a a a 1 1 a r; , 3 0 1 2 , ’ , ’ 0 3 i ’ a , 61 3 b3 i “ z z i b2 “ 0 2 . cos ( i k ) cos (j k ) ( 4 8 ) cos (k k) , j+ ’ i + bl 2 2 a ’ , ’ ’ 1 j 2 2 s ’ ba ’ j “ 2 3 O3 C 3 2 2 c l 0 2 o s . k ’ k ’ k ’ VE CTOR ANAL YSIS 1 06 the p d n e d rection but to lie wholly in the la an from i i -plan e k form upon the left it is seen to lie in the it s Hence it must be the l ine of intersec t ion of those two planes / T h s gives the or 1 i s u magnit de Its m m scalar relation s k j j ’ ’ . i . — a magni t ude 1 a l is the square of the sine of Hence the vector b etween t h e vectors i and i z Th e th e angle ’ . — “ k 2 j 53 ( ) is the line of inte rsection of the k and jk-plan e s and its magni tude is the sine of the angle be t ween the planes Eight other s imilar vecto rs may b e foun d each of whic h gives one of the n ine lines of inte rsection Of the t wo sets of m u t u a ll y ort hogonal planes Th e magn i tude of the vector s in each case the sine of the angle b etween the pla nes Various ex amples in Plane and Solid G eometr ma be solved by means of products Exa mp le 1 Th e perpendicul ars from the vertices of a tri a n gle to the opposite sides meet in a point et A B C b e the t rian gle et the perpendiculars from A to B C and from B to CA meet in the point 0 To Show 0 0 is perpen di cular to A B Choose 0 as ori gin and let OA A O B = B and 00 Then C ’ ’ , . , i y y . . . L . . L . . , , . By y B C= C— B, C A = A — C, AB z B — A . h po thesis A and Subtra ct ; (C B(A C(B B) O C) 0 A) 0, . which proves the theorem Exa mp le 2 : To fin d the vector equation of a through the point B parallel to a given vecto r A . . drawn K W PR O DIRECT AND Let S E 0 be D UC TS OF VE C TOR S 1 07 L the origin and B the vector O B et B be the ra di ns vecto r from O to an point of the required line Then B B is parallel to A Hence the vector p ro du c t va n is h e s y . . . . q Thi s . AX ( R — =0 B q . A is the desired e uation It is a vector e uati on in the unkn own vector B Th e e uati on of a plan e wa s seen ( page 8 8) to be a scalar e uation such as . q q . 3 0 c 0 in the unkno wn vecto r R Th e point of intersection of a line an found at once Th e equations are . d a plan e ma y be . B R C A> <R 0 ) c A> <B > <C ACB CRA ACB c A c Hence A i O In th s cas e h o w solution evidently fails when A C ever t h e line is parallel t o the plane and there is no solution ; or if it lie s in the plane there are an in fi nite number of solu ti ons E xa mp le 3 Th e introduction of vectors to represent planes Heretofore vectors have b e en used to denote plane a re a S o f de fi nite ex tent Th e direction of the vector was normal to the plane and the magnitude was equal to the area to be re presented B u t it is possible to use vectors to denote not a plane area but the entire plane itself just as a vector represents a point Th e result is analogous to th e p la n e co ord nates of analytic geometr et 0 be an assumed origin et M N be a plane in space Th e plan e M N s to be denoted b a vector Th e . , , . . ' . . i , y L . . . i . L y VECTOR A NAL YSIS 1 08 p whos e di rection is the direction of the perpendicular drop ed upon t h e plane from the origin 0 and whose magni t u de is the Thus the nearer r e c ip ro c a l of t h e length of t hat perpendicular a plane is to t h e origin the longer will be the vecto r whi ch represent s it If r b e any radius vector dr awn from the o ri gin to a point in t h e plane and if p be the vector whi ch denot e s the plane t hen . . , t is the equation of the plane o 1 p Fo r . cos ( r p ) p N o w p t h e lengt h of p is t h e reciprocal of the perpendicular distance from O to t h e plane O n the other hand r cos ( r p ) is t hat perpendi cul ar dis t ance Hence rp mus t be unit Ifr and p b e e x pressed in terms of i k t o r p , . , y . . , r = = p Hence ro p = x u xu i + y j+ i + v f + y c z j , . , k j+ w k + z w l z quantities u r w are the reciprocals of the intercepts of the plane p upon the ax es Th e relation bet ween r and p is symmetrical It is a rela tion of duality If in the equat ion Th e , , . . . ro p = 1 be regarded as variable t h e equation represents a plane p which is t h e loc us of all points given b y r If however p be regarded as variab le and r as constant t h e equation re p re sents a point r through which all the planes p pass Th e development of the idea of duality will not be carried out It is familiar to all students of geometry Th e use of vec to rs to denote planes will scarcely be alluded to again until Chapter VII r , . , . . . . VE C TOR ANA L YSIS 110 i vanishes is t hat their vector product van shes mutative laws do not hold . Th e com . AXB = i> <i B XA > <j j i xj AXB ( A2 B S k xk jx i k xk j k> <j= i k> <i i Xk = As 8 i 2) A B ( s ( A1 AXB j A1 B l B 3 A2 3 i j AI A2 A B 3 B ] 0 a )j k 1) k 2 Th e q l scalar triple product of three vecto rs [ A B C] is e ua to t h e volume of the parallelopiped of wh ch A B C are three edges whi ch meet in a point i , . [ A B C] AB xC B Cx A CAxB AxB C B xCA CxAB [ A B C] o [A CB ] . , DIRECT If the KE W PR ODU AN D S T VE C TOR S C S OF component of B perpendicular to A be 111 ” B , AA CB A- AB C ACB CB A CB D A- ADBC o C D] B [A C D] B [ D D] 0 A [ [ A B C] D A ( 26 ) . q e uation which subs ists between four vectors Th e A, B , C, D is [B C D] A Th e ( 27) . ication of formul of vector analysis to obt i the for of Plane and pherical rigonometry system of vectors is said to be reciprocal to A ppl mul a = 0 A B C 1) [ ] [D A B ] 0 C [ D A] B a n a T S a ’ , b ’ , c . ’ the sys t em of three non-coplanar vectors b x when [ b a c x a [a b c c ] , r b, a, c ] C a x b [a b 2 9 ( ) vector r may be expressed in te ms of a set of vectors and its reciprocal in two similar ways A r r a . r y ’ a ro a , a , b b r c bb r e r . ’ ’ r . ’ ’ 3 ( 0) c y ( 31 ) necessar and su ffi cient conditions that the two s stems of non-c oplanar vecto s a b c and a b c be reciprocals is that Th e r ’ , a a ’o b a ’ o c y l . , , b b b 1 o c 2 “ r Ma c form a s stem eciprocal to form a s stem reciprocal to a b c Ifa ’ , b ’ , y 0 ’ ’ , [ a t b l c ’ , ’ a ’o ’ ’ ] ’ , ’ . a, ’ o b b, C ; 2 3 ( ) 0 . then a, b, c il w l VECTOR ANAL YSIS 11 2 P A B P- C P- A Q- AB a c BA AB B C o j system i k is its own reciprocal and if conversely a syste m b e its own reciprocal it mus t b e a ri ght or left han ded syst em of thr ee mutuall y perpendi cular unit vecto rs App l i cat ion o f t h e theory of reciprocal syst ems to the solution of scalar and vecto r equati ons of t h e fi rst degree in an un kn own vector Th e vector equat ion of a plane is Th e , , . . r A a ( 36 ) . Applications of the methods developed in Chapter IL to the t reat men t O f a system of forces acting on a rigid body and in part icular t o the reduction of any system of forces to a single force and a couple of which the plane is perpendicular to that force Application of the methods to the treatment of ins t antaneous motion of a rigid b ody o b taining , . (46) where v is the velocity of any point v a translational ve l o c ity in the direction a and a the vector angular velocity of ro t a t io n Further application of t h e methods to Ob tain the condi t ions for equilib rium b y making use of t h e principle of virt ual velocities A pplications of the method to ob tain t h e rela t ions which e x ist b etween the nine di rection cosines of the angles b e t ween two systems of mutua lly orthogon a l axes A pplicat ion to special prob lems in geome t r including t h e form under whi ch plane co o rdinates make their appear ance in vector analysis and the method by which planes ( a s , 0 , . . . y 1 14 VE C TOR ' y AIVA L YSIS b vecto r methods that the formu ume of a tetrahedron whose four vertices are 12 . Show v ( “ v y rs 2 1 ( x2 ) , z 312 : “ H 8 6 a x ( 3 z) I ’ I ll z N 3 5 z y3 z 2 s “4 z 75 , l a “ ) ( 3 p for the 90 z l ) 1 , 1 2 1 8 1 , use of formula ( 34 ) of the te xt Show that [ where a , b, a b c ] = a bc are the lengths of c a , n 1 l m l 1 i y respect vel b, c an d where cos ( b c ) m cos ( c a ) n cos ( a b ) uantit ) I4 Determ ine the perpendi cul ar ( as a vector which is dropped from the origin upon a plane determined b the termini of the vecto s a b c Us e the method of solution given in Art 4 6 1 5 Show that the volum e of a tetrahe dr on is e ual to one sixth of the product of two opposite edges by the p e rp e n di c u lar distance between them and the sine of the included angle 1 6 If a line is drawn in each f a ce plane of an t ri e dra l angl e through the verte x and perpen di cular to the thi rd e ge t h e t hree lines thus obtained lie n a plane t : , , , , , A r . , , q y y . . q . . y . i . . d , C H AP TER III TH E DIFFER E N TIA L CA LCULU S OF VE CTOR S Dife r e n ti a tio n f o Fun c tio ns f O On e Sc a la r Va ri a ble a vecto r varies and changes from r to r the in cre ment of r will be the d fference between r and r and will be denoted as us u al b A r IF y ’ i ’ . Ar = r r y where A must be a vector quantit If the variable T b e unrestri cted the increment A r is of course also unrestri cted it may have an magnitude and an direction If however the ve c t o r r be rega rded as a function ( a vector function) of a Single scalar variable t the value of A r will be completely determined when the two values t and t of t which give the two values r and r are known To o b tain a clearer conception of the quantiti es involved it will be advantageous to thi nk of the vector r as drawn from a fi x ed origin 0 ( Fig When the independent variable t changes its value the vector r will change and as t possesses one degree of freedom r will var in such a way that its te rminus describes a cu ve in space r wil l be the rad us vector of one point P of the curve ; r of a neighboring point P A be the cho d P P of the curve Th e ratio y . y , . ’ , ’ . , . , y r i r . ’ ’ , ’ . Ar A t r , VECTOR ANAL YSIS 116 will b e a vector collinear with the chord P P b u t magnifi ed in t h e ra ti o 1 : A t When A t approaches z ero P will a p proach P t h e chord P P will approach t h e tan gent at P and the vector 03r will approach ’ ’ . ’ , , dt which is a vec t or tangent to the curve at P di rected in that s ense in which t h e varia b le t increases along the curve Ifr b e e xpressed in terms of i k as j , the components ( r r 1 , r Ar = r , r = r 2, r 3 will be functions of the scalar t ’ — Ar ( r2 j 2 r 3 k A r2 ) j A r3 ) k ( r3 = A r i + Ar 1 2 j+ A r3 k r A , A t dr r . A rl At i + 1 A r1 ) i + 1 . r A 2 A t d rl d rg , d t r 3 k A t , d t d r 3 d t k ( 2) ' Hence the components of the fi rst derivative of r with re spect t o t are the fi rst derivatives with respect to t of the components of r Th e same is true for the secon d and hi gher derivati ves . . d z dt d " r ‘ i d r 2 d t r d t” In d “ 1 2 r d t " s rmrla r a manner if r non-coplanar vecto rs a b , ' 1 1 , e xpressed in terms as be c a a + bb + c c y an three VE C TOR 118 A(a Ab b) o At Hence in the limit when 3 7 A + 0, At ( ( a o b x c ) a 8 ( da X -b x o a Ab At At db X b)= Aa Aa d a NAL YSIS b x c . x x[ x b c ] ( 6) . d y last three of these formul a may be emonstra ted exactl as the fi rst was Th e formal process of di fferentiation in vector anal s is differs in no way from that in scalar analysis e x cept in this one point in whi ch vector analysis always di ffe s from scalar anal sis namely : Th e order of the factors in a vector product Th e y . y r , THE DIFFERENTIAL CAL CUL US OF cannot be cha nged wi thout changing the sign Hence of the two formul a ECTOR S 119 V Of the product . d d r ly the fi st is evident incorrect but the second correct In other words scalar di fferentiat ion must take place w thout altering the order of the factors of a vector product Th e factors must be fferentiated in s i tu Thi s of cou se was to be e xpected In case the vectors depend upon more than one variable th e results are practicall the same In place of tota l deriva t ive s with espect to the scalar vari ables partial derivatives occur Suppose a and b are two vectors which depend on three scalar variables x y z Th e scalar product a h will depend upon these three variables and it will have th ee a t al der vati ves of the fi rst order i , , di r . . . . y r . , . , pri , o . i , . ( a o h) o b + a o b + a o b + a o o . ri second pa t al de ri vatives are formed in the same 9 2 a wa y as 9x r VECTOR ANAL YSIS 1 20 O ft en it is more convenient to use not the derivatives but t h e di f feren t ials Thi s is part icularly true when dea ng wi th rs t differentials Th e form u l a (4 ) b ecome fi li . . d ( a o b) = da -b + a o db , and so forth As an illus tration consider the following If r be a un i t vector e xample . . r o r l z . Th e locus of the te rminus of r is a spherical surface of uni t radius described about the origin r depends upon two vari ab les Different ate t h e equation i . ( d r) . . o r Hence + = 0 r . r o dr z 0 . Hence the incremen t d r of a u n i t vecto r is perpendicular t o t h e vecto r Thi s can b e seen geometrically If r traces a sphere t h e variation d r must b e at each point in the tangent plane and hence perpendi cular to r V ector methods may b e employed advantageously d in the iscuss ion of curvature and torsion of curves et r deno t e the radi us vector of a curve . . . L . where f is some vector function of the scalar t In most a p p li ca t ions in physics and mechanics t represents the time et s be t h e lengt h of a re measured from some de fi nite point of the curve as origin Th e incremen t A r is the chord of t h e curve Hence A r A s is appro ximately equal in magnitude t o unity and approaches unity as its limi t when A 3 becomes in fi nites imal . . . . . L VE C TOR 1 22 t d A NAL YSIS t = e -c = n o =n =1 n o t =0 an Different at ng the fi rst set dn =0 dc =n t dt =c and the second d n + dc c c t d c t d + B u t d t is para e to c and consequentl y pe rpend cula t ii C o o n o o . o , ll l i r o o n dt = o . to n . . t = 0 dn Hence B u t the ncrement Th e increment of n s pe rpendicular to t of n is also perpendi cular to n It s therefo e parallel to e d u /d s it is pa a el t o d n and hence As the tortuosit is T to c Th e tortu osit T is i y y . d dc d z a s r ll X d x a ) " z s 1 h / d x s m d x d s s z + d x s d . 1 z 1 3 r d dr 2 r d dr + ds ds r 3 r d dr d r . , r ? i . t ( i . ds 3 Vc . c 1 d s x/ c Th e fi rst te rm of thi s expression v a n ishes T moreover ha s 2 b een seen to be para llel to C d r/d Conse uentl the magnitude of T is the scalar product of T b the un it vec tor c in the direction of C It is desirable however to have the to rtuosit positive when the normal 11 appea s to turn in the posit ive or counte rclockwise direction if viewed from that side of the n o-plane upon w h ich t or the positive part of the curve lies With this convention d 11 appears to move in the direction — c when the tortuosit is positi ve that is 11 turns away from C Th e scala va ue Of the tortuosit w therefore be given by — c T q . y y y r . . r l . . y , y ill , THE DIFFERENTIAL B ds x d s 1 d dr m 3 is parallel to the vector ut c r > < da An d AL C UL US OF VE C TOR S 3 r d dr T — c C d x ds al d z d 3 / s r d z ds Henc e 2 r d z 1 23 . z r — 2 s T ( is a u it vector in the direction H e c c C n d C Hence n e . d T 2 2 d dr d ds d d 2 2 d 3 3 1 r dr r d8 2 s x d 3 3 1 C C r 2 r z r d 2 2 d r iy y i r q d3 r 3 y r d i to tuos t ma be obta n ed b another metho wh ch s somewhat sho ter if not uite so st aightforward u n t t = 0 C = c — d c dt c t ence — dn dc u c — dt n t dn - i H Th e o . o . o o o . l H is parallel to c ; hence perpendicu ar to 11 ence 0 B u t d n s perpendi cular to 11 0 Hence d n t dt n Hence d 11 must be parallel to c Th e to rtuosit is the mag of d n d 3 taken however w th the negative sign n it u de because d n appea s clockwise from the positive d rect on of the cu ve ence the scalar tortuosit T ma be g ven by NO W d t i . . . r . H r dn T=t > < i y dc 0 . y . i i y i 1 4 ( ) VECTOR A NA L YSIS 1 24 d C t x c d CC— t x c -C / C C x . s t x c -C = 0 d C t x c t C dr d s d 2 ds Cartesian -C dr 2 d 2 r 2 s d 3 d 2 d d s r s 3 r 2 coordi nates this b ecomes d x ds d dy ds ds m d s d 3 d s z a 2 x 3 ol s 2 d 3 d d y 3 ds d s 3 s 2 e 3 Th os e who would pursue the study of twiste d surfaces in Space further from the s t andpoint of vecto rs w l fi n d t h e boo k Ap p lic a tio n de l a M éth o de Ve c to rie lle ole G ra ss m a n n a l a G eo metri c In n ités i m a l e 1 b e xtreme F E H B fi y il ly “ ‘ ” 1 Pa ri s , Ca rré e t Na u d , 1 89 9 . VE C TOR A NAL YSIS 1 26 i d introduced b y N ewton It w ll als o be conve n ien t to enote the unit tangent to the curve b y t Th e e uat ons become . q i . is the rate of change is a vector uantit et it be denoted defi ni tion Th e a cc e le ra q ti o n y L . Av LIM M " _ i Df ferentiate dv G At i dt dv d dr dt dt ll t the expression dv — dt d — r — v ( v t) d dt dt dv dt z d i s dt 2 s dt dt d dt ds d t r r where C is the (vector) curvat u e of the cu ve and v is the speed in the curve Substituting these values in t h e e uat on the result is d “ b q i . “ 7 a a A= é ? Th e . s 2 t + v C i . “ i f “ t” i acceleration of a particle mov ng in a curve has there fore been broken up into two components of which one is p a ra l le l to the tan gent t and of which the other is pa allel to the curvat ure C that is p e rp e n dic u la r to th e tangent That thi s resoluti on h a s been accomplished would be unimportant were r , , . TH E DIFFERENTIA L CAL C UL US OF VE C TOR S 1 27 it not for the remarkable fact which it brings to light Th e component of the acceleration parallel to the tan gent is e ual in magni tude to the rate Of change of sp e e d It is entirel ha t sort t ll rve the p a rti cle s describing indep e n de n t p f w It would b e the same if t h e particle described a right line with the same speed as it describes th e curve On the other hand the component of the a cceleration normal to th e tan gent is equal in magnitude to the product of the s uare of the Speed of the particle and the curvatu re of the curve Th e sharper the curve the g eater this component Th e greater the Speed of the particle the greater the component B u t the rate of change of sp e e d in path ha s no e ffect at all on this normal component of the accelera ti on k as If r be e xpres sed in te r m s of i . i . “ V q . [ . q , y . . q , r . . , . , x A = v= i + y , j+ 515 1 + y r A = v= j . z k, j+ z k, + y y + x x s z z From th ese formul a the difi e re n c e between 3 the rate of change of spe ed and A i the rate of change of velocity is apparent J ust when this difi e re n c e fi rst b ecame clearly recogniz ed woul d be hard to say B u t certa in it is that N ewton must have h a d it in mi nd when he stated his secon law of moti on Th e rate of change of velocit is proporti ona to the impressed force ; but rate of change of speed is not Th e h o do gra p h was introduced by Hamil ton as an aid to the s t udy of the curv li near motion of a particle With any a s sumed origin t h e vector velocit i is laid o ff Th e locus of its term inus is the hodograph In other words the ra d us vector in the hodograph gives the velocity of the , i , , , . d l . y . . i i y . . . , VECTOR ANAL YSIS 1 28 y particle in magnitude and dir ection at an instant It is pos s ib le to proceed one step further and co n struct the hodo graph o f the hodograph This is done by la ing OR the vector acceleration A r from an assumed origin Th e radius vecto r in the hodograph of the hodograph therefore gives t h e accelerat ion at each instant 2 F 9 i et a particle revolve in a circle e 1 m l x a E ) ( g p of radi us r with a un iform V a n g u l a r v e l o c i t a Th e speed of the particle w ll then b e equal to . ‘ y L . . y L et v a . i r. the radi us vector F G 29 drawn to the particle Th e is perpendi cular to r and to a It is i v a >< r I he r . . . velocity v . . vector v is always perpen di cul ar and of cons tant magni a r t ude Th e hodograph is t herefore a circle of radius v Th e radi us ve c t o r i in this circle is just nine t y degrees in advance of the radius vector r in its circle and it c o n s e quen t ly desc ri bes t h e circle with the same angular velocity Th e acceleration A whi ch is the rate of change of v is a always perpend cular t o v and equal in mag n itude to Th e . . ' , i A = Th e accelera tion T : as Hence B ut a A o a ff v a z r . may b e given by the formula A xv a a ct a a oa r . is perpendicular to the plane in whi ch i z Th e A - — a a r —a 2 r r lies a r , 0 . . accelerat ion due to the uniform moti on of a particle in a circ le is directed toward the centre and is equal in magni tude to t h e square of the angular velocity multiplied b y the radi us of the circle . 1 30 VE C TOR l A NAL YSIS lly Perhaps it wou d be well to go a li ttle more ca refu into th s question If r be the ra di us vecto r of the pa rticle in its pat h at one instant the ad u s vector at the ne x t ins ta nt Ar Th e area of the vector of which r an r is r A a re t h e b ounding ra di i is appro ximatel e ual to the are a of t h e A a n d th e triangle enc osed b chord A r This a rea is i r i . , l yr r r , 1 d yq . , . 1 1 r Th e rate of consequently l l LE I At r 02 ’ - description of are by r) 1 a ' ra the Ac e s o di us £ 3: z A , is B 4 r . vecto L ‘ r is t v r " 5 L et Tand T be two values of the velocity at poin and which near together acce l e rati on at the limit of t wo o P x A ” LIM At r Po a re re ak up the vector r l Th e . A l E i r r° A t into p i lr i io components t wo pa alle and the other per end cu a to the ts P0 one a c c e l e ra t ro n ro Af A t y _ x l/ n? o. Th e unit when A t approaches z ero approaches z ero when A t approaches z ero r p oac he s q antity qu tity u Th e . . 2: an a p y DIFFEREN TIAL TH E Hence r A L C UL US OF VE C TOR S C 1 31 A x r — rO X ro r i i ch of the thr ee te ms upon the ri ght-hand s de s an in fi ni tesimal of the s e c o n d order Hence the rates of de s c rip ti on of a ea at P and P difi e r b an in fi nitesimal of the Th s is true for an s econd order w th respect to the ti me point of th e curve ence the rates mu st b e e xactl e ua at all points This proves the theorem Th e mo t ion of a rigid bod one po nt of wh ich is fi x ed is at an instant a rotation about a n i n sta ntaneous a x is passing through the fi x ed po int et i k be three axes fi x ed in the bod but mov ng in spa c e et the ra di us vector r b e drawn f om the fi x e point to an point of the bod Then B ut e a r y . i o H . y . y L j L y , y d d Substit uting r q d . x r r i + y o j+ z d i + y d j+ k, dh z r i, d rj r , d k z w ( i o di + y i o d j+ z i + ( j o d i + yj o d j+ z j + (x k x i i i d = (d r ut ce y r the values of d s econd e u ation H en i . , x B y y q l . . th e i . d j+ o = j j j o di =0 = d k k d 0 + j j k di + i dk = 0 r di = o o or j-d i or k -d j or i d k o = d k dh j i =o . obtaine d k) i d k) j d 1 32 VECTOR A NAL YSIS i i r d x u st t t ng th s values in the e pression for e e u S b — — d i z k d r = (z i d k yj — x i k d k d )k + (y j is Th o is a vector product . ~ d i) i . . d k j+ j a =k dj di , " o d t d t d r d t k . > <r a d t 3 . Thi s dy i li shows that the instantaneous motion of the b o s one a about t h e ne a Of rotation with the an gu lar velocit Th e Thi s angular velocit chan ges from n s tant to instant proof of this theorem fi ls the lacuna in the work in Art 5 1 Two in fi nitesimal rotati ons may be added ike vecto s et a , and n a be t wo angular velocities Th e d spl ac e ments due t o them are y i l y . L l i . Ifr be d splaced by i r If it a, d1 r d2 r = r y then be displaced b a z r , + a x l r d t . d i r ( d t ) of or er h ghe tha n the 2 fi rt s , i d r = a l . it becomes Hence neglected r . . it becomes + d1 If the in fi n ite s im a ls . > <r d t + r x a 2 wh ch proves the theo em . Ifboth a 1 ' 1 ‘ 3 r d t, i id d y sides be d v e b 2) x r. dt be 1 34 VECTOR A NAL YSIS r l i r i y whe e C is some co nstant vector To a ccomp is h th e nte g a tion in a n y pa rticular cas e ma be a m att er of some d ffi cult just a s it is in the case of ordi nar integration of sca ars Integrate the e u at on of mot on of a Exa mp le 1 rojectil e Th e equat on of motion is s im l y p y q i py i . . g, r r i y a wa ve r . f= g t + b , whe re b is a constant of integration velocit at the time t 0 y . l i i l ys which exp e s ses th e fact that the acce erat on s ticall do wnward and due to gravi t y l It . is evi dently th e . i i is a nother cons tant of integrati on It s the posit on vecto r of the point at time t = 0 Th e path wh ch is given b th s as t e uation is a parabola That th is is so ma be seen b e xpress ing it in terms of x a n d y and eliminating t Th e rate of description of areas whe n a a Exa mp le 2 t o le moves under a central a ccelerati on s cons t ant f ( r) Since the acce erat on is parallel to the a us c . l q . i y . . i i l i r H r _ — - ci — a nd which > < d d ence r proves th e t tement s a r di . x . T : C, 0 , y i y pr TH E DIFFERENTIA L L UL US CA C q Integrate VE C TOR S OF 1 35 ri the e u ati on of motion for a pa t cle moving with an accelerati on to ward the cent e and e ual to a constant multiple of the in ve se s ua e Of the stan ce from the centre Exa mp le 3 r r q r q di . G iven Then Henc 1 ' X rx e e 0 r i . C . q u tio s together with x rx( x ) { r r a n . 1 r r 7 r 3 r -} r r r r . o H ence Ea c h s id e In te grate . r i i yr i q ality is a perfect of th s e u difi e re n tia l . Then i iy whe e e I s the vector constant of inte grati on e s its ma gni tude and I a unit vector in its rection Mul t pl the e ua t on b di r x i 0 2 C r . . r + e r o I . q VE C TOR 1 36 A NA L YSIS and cos = r + p y cit . r , I) . e r c os u . P r This ( c os it : 1 + q i e c os u i dir i is the e uat on of the ellips e of wh ch e IS the e c c e n tri ect on of the major Th e vector I is drawn in the Th e length of th s ax is is i P a 1 e 2 i is p o ssible to carry the integrat on further the time So far merel the path has been found It y . Sc a la r Fu n c tio n s and obt ain . n i n Sp a P s i t i o o f o ce Th e Op e ra tor V . lr fun ction V (x y z ) which takes on a de fi nite sca a value for e a ch set of coordinates x y z in space s called a scalar function of position in space Such a funct on for e x A , i i , , , . a m p l e , is x 2 + y 2 + z 2 = , r is function gives the s quare of the d st nce of the point Th i function V will be supposed to be in general continuous and single -valued In physics scalar fun ctions of position are of constant occurrence In the theor of heat the te mperature T at any poin t of a body is a scalar function of the position of that point In mechani cs and theories of attraction the potential s the all-importan t function This too is a scalar function of position If a scalar fu nction V b e set equal to a constan t t h e e ua tion (x , y, 2 ) from the origin a . Th e . y . i . , . , . , V ( x, y , z ) c . q ( 2 0) de fi nes a surface in space such that at every point of it the function V ha s the same value 0 In case V b e the tem era . p VE C TOR 1 38 A NAL YSIS r t i cre s vector sum which is the re s u l t a irt o f V is denoted b V V Th e y a e of n a e . VV i 9 9V 9V 9V — 9 9 g x " 21 o z dir cted rate of cha ge of i d re cted or vector derivat ive of V so to speak Fo r th s reason V V will be c a lled the de riva tive of V; and V the p ri m i tive of V V Th e te ms gr a die n t and s lop e of V are a so us ed for V V It is custo mar to egard V as an operator which obta ns ro m a scala r functi on V of posit on in space a vecto V V f V V re p re s e n ts n e a r a l , y r r . i . , . V i 9 _ i 9 9 31 x 9 i 9 a V=i i is symbolic perator ( 22 ) . 9 9g x . z y was introduced b Sir W R amilton and is now in uni versal emplo ment There 1 seems however to be no uni versally recogniz ed name for it although owing to t h e frequent occurrence of the s mbol some name is a pra ctical necessit It has been found by experience that the monos ll able de l is so short and eas to pronounce tha t even in complicated formul a in which V occurs a num ber of times no inconveni ence to th e speaker or hearer arises from the repetition V V is read Simpl as del V Although th s operator V h a s been de fi ned as H Th V O y y y i V m p ha r i co ne th e te r e us e Ot h e rs h a ve the no ne t h e wo r e tric itcit a vo ds ‘ na m e is y . y — Ij TE C ” 9 ay a ny s ea , e th n te . o wn it s n o te ec a is t o b e y so bla t o its e n e ss e a nc e to an A tle d re s e n ve rt e a nc e d Aand to a n h a ve de s is re a b ut fo r na t no t e c tu r n to o t h a t oe s o n a n d re d vul a nd no t e e rs t o . In ur os e s o c o n us e the ee f , s as “ r nt n n s t ru c t o n th e s e a er or ue n t e t t e rs h e o rie de r o fo r Ass yr a n c o ns e t h e o r e r o ft h e n ve rt n Flip p l in h is E infi h rung in die M a xwe ll ’s c h e . s a t o o e u h o n o us n a n e c e s sa r — q i d o ft e n re u re lta H o w th V ” . i , m Nd liki g f i i d mbl i q l d p i m by i i g d l dD T E i p i l ig i f ymb l O i d i gd d d pi i g p i l l i g p p i i m i g m i g d f p k p d So . y . 9 . . , , 1 . so y in le o die p e ra t ion no a rt c u a r e b e a re r th n e ve n rs re wh e n TH E DIFFERENTIA L C AL C UL US OF VECTOR S 1 39 so that it appears to depend upon the choice of the axes it is in reali t independent of them This would be surmised from the interpretation of V as the magnitude and direction of the most rapid increase of V To demons t rate the inde ’ ’ k and a new set of n c e tak e another set of a x es i n d e e p ’ ’ eferred to them Then V referred to th s variables x y z s stem is y , . . ’ y , , r V I , , j ’ i , . v — ‘ 9 9 e 9g ' x 9 + k ’ 9 z r ( 2 2) ’ i and from i j to j ctually carr ing out the erentiat ons and taking into account identities ( ) an d act ally be transformed into B y mak ng use of the formul ae 1 04, for transformation of a x es a , di f f y , k i ’ , ’ , i the 49 V u V ’ page k and b b fi nall ’ V ma Art 53 , ’ y y y y . V r . deta ils of the proof are omitte d he e be caus e s horter method of demonstration is to be g ven Consider t wo surfaces ( Fig 3 0) Th e , i an othe r . . V ( x, y , z ) = c and i i y upon whi ch V is constant and which are more over nfi n tel near together et x y z be a given point upon the surface V 0 et r denote the ra di n s vector drawn to thi s d ’ ‘ T poin t from an fi x ed origin V(f iJ VW Then any point near by in g the neighbo ring s u rface V c d c may b e represente d by t h e radius vector r d r Th e a c tua l increase of V fro m t h e fi rst surface to the second is a fi x ed uan ti t d o Th e ra te of in c eas e is a va r a bl e : . L L . , , y . ‘ c . q y . r i VECTOR A NAL YSIS 1 40 quantity depends upon the direction wh ch is fol lowed when pa sing from one surface to the other i dr and s . L et be a u it normal to the segment of that normal intercepted wi ll then be the least value for di rection d r surfaces and d rt the between the surfaces Th e quotient dr , n n . d ri 11 . d 0 i r ll l will therefore be a maximum when d r s p a a e to e ual in magn tude of d n Th e e xpress on q i i . n an d 23 ( ) n di dr is therefore a vector of whi ch the rection is the i ection of most rapid increase of V and of whi ch the m a gn it u de i s t h e rate of that increase This vector is entirel independent of the axes X Y Z et d c be replaced b y its equal d V which is the increment of V in passing from the fi rst surfa ce to t h e second Then let V Vb e de n e d again as y L . . , , fi . 1 VV k w». From dV n a 24 ( ) . this de fi nition V V is certai nl y the vector whi ch give s the di rection of most rapid increase of V and the rate in t hat direc t ion Moreover V V is independent of the axes , . . V Vdr fn dI — d o 2 ( 5) dr . n is a un it normal Hence n d r is the projection of d r o n n and mus t b e equal to the perpend cular distance d 77 be t ween t h e s urfa ces n . . i . 1 42 VECTOR ANAL YSIS q A? i r Moreover thi s e uati on de fi nes dv dz In a s milar m anner it is possib le to lay down the following defi nition function of Defin itio n : Th e derivative V V of a scal a po sition in space s h a satis fy the e uation . ll d q r VV . dV for all values of d r mpo rta n t This de fi n iti on is cert ainl the most natural an from theoretical considerations B u t for practical purposes either of the defi ni ti ons before given seems to b e bett er Th e real sign ifi canc e of th s last They are more tangible de fi ni t ion c a nnot be app e ciate d until the subject of linea vector functions has been treated Se e Chapte r VII Th e computation of t h e derivative V of a fun ction is most frequently carrie on b means of t h e or d nar dif ferentiation y . di . . i r . d . r . y i y . Let 9 Vr =i W 9 3/ x x H V W Oz y V z Hen 1 (1 x + jy ce kz ) Vr Th e dir derivative of r is a unit vecto r in t h e e ction of r This is e videntl the di rection o f m o s t ra pid in c re a s e of r a n d the ate of tha t in c ease r y r . . TH E D IFFEREN TIAL A L C UL US C VECTOR S OF 1 43 V ( x, y, z ) j a: 1 ( 23 2 y 2 z k He n c e V q c s ‘ . 1n ua re y z ( 2 90 ” v J w— — l — k y z ) 1 i a vector whos e di recti on s tha t and who se magn i tude is e ua to the recipro c a of th e of th e length r l /r is q l l . VT “ Th e ) Z r derivative of , i 1 V of - r 2 y r r p oof is left to the ea der Exa mp le A: Let r n r . log m V( co, y, z ) + y J2 + 33 y ” 0k 1 x If ( cc , y, r denote the vector + y 2 from the origin z ) of spa c e the fun cti on V m a y be wr t te n dra wn i , i H 2 x + = jy r r — e n 00 - r r ( r— o kk r 3 r (k ) r— — k kd r) k k- o -( r — k k r) o t o th e as pit o n VE C TOR 1 44 There A N AL YSIS is another method of computing upon the identit VV= dV Ci r - i wh ch V y . — dr Hence L et V: dr VV o . v r where a is a, d V= dr o Exa mp le 3 : r constant vecto s L et o 2 dr o Henc e ( rx s a V= da a r a o . ) ( r xb o a ( o h . dr b ro b a a ro b o ra ) , W here - V V= 2 r a h V V= . ro r a o h — r a ro'b ( ’ w v a r Ar a de PN dr h . . . Ar m dV V: r = dr V V a V V= Hence consta n t ve c to a a ro h ) ( a x b ra ( o A fi r “ - A d b " dr V V r a - r a b h rxb ; an a -) r a ) . Which of these two methods for computing l shal b e applied in a particular case depends enti rely upon their relative ease of execution in that case Th e latter method is independent of t h e co ordinate ax es and ma therefore be preferred It is als o shorter in case th e funct on V e an be expressed easily in te rms of B u t when V can not be s o e xpressed the former method has to be resorted to Th e great importance of the ope rator V in mathe me ti cal physics ma be seen from a few illustrations Su p P08 9 T ( w 31 z ) be the tempe atur e at the point a y z of a . r . y V i . . y , , r . , , 1 46 VE C TOR A NAL YSIS il y l y Pote nt a in el e c tricit or magnetism is the pote nti a energ per un it charge or pole ; and pote ntial in attracti on problems is potent ia l energ per unit m ass ta ken however w th t h e negative sign It is ofte n convenient to t eat a n op era to r a s a uan tit provided it obe s t h e same formal laws a s tha t f fe e n t a to s ua n tit Cons ider for e x ample the a t a y , y y y . 9 9 i r p r i l di r i r . q q , 9 9 9 3/ x 9 z d r l d di r far a s combinati ons of these a re c oncern e t h e fo mal a ws a re precisely what the woul b e f in s tea of ffe entiators three true scalars As y d i a r i b, , , c we e g ven Fo r insta nce t h e commutativ e l a w . 9 th e as s 9 i oc ati ve x 9 9 9v a 9g b =ba, la w 9 — a nd i 9 3: 9y 9 9 9 z (b c ) the d s tributive law 9 9 92 a 9 9y x + 9 9 x 9 9 y r + 9 x 9 z l hold for the differentia to s just as for sca ars formu a a s l . Of cou rs e such DIFFERENTIA L TH E the same oo k ng upon In l i v e c tor It is . S VECTOR OF y 1 47 i d y a great advantage ma be obta n e b wa y a _ i V as a CALC UL us 9 cc a a 9y 9 r z i not a true vector fo the c o e ffi c e n ts , 9 l 9z 9 3/ 3: i are not true sca ars It is a vector d fferentiator and of cou se an operand is alwa s implied w th it As far as form a l opera t ions a re concerned it be haves like a vector Fo r ns tan c e r y . i i . . = ) V n + V v, V (n V ( a v ( V ie ) v ) v e V V n ( ( V v) n e n ) , , if and are any scalar f nctions of the scalar variables and if a scalar independent of the var ables with regard to which the i erentiations p rformed rep resent any vector the formal comb nati o x, y t wo v n , z e u i be df f e a re . i If A A V is o A V ° r i his operator — Al 9 am + A2 9 9 3/ A = Al i + Ag j p o v ded Aa k . is a scalar di fferentiator When to a scala r function V ( a y z ) it gives a scalar T A V . , . , 9 V = A1 V V A ( ) . Suppos e 9 x for convenience that ( V) a V= A 7 1 9 - da 1 + Az 9 V ay +4 9 V , . 9 z is a unit vector 9 V —— ; ag a 9 V + a 3 az . n VE C TOR A N ALYSIS 1 48 li a h t e the rection c sine s of ne referred o d i where l to the axes X Y Z Cons equentl ( a V) V appe a rs as the well -known direc tio n a l derivative of V in the d rection a Th s is ofte n written a , y “ r a s a re . , , i i WM 9V 7 1 9 S . Z e xpresses the magnitude of the rate of increase of V in the direction a In the particular case where this direct on 1 8 the normal 11 to a surface of constant value of V thi s relation becomes the normal derivative i It . , . ( ” W V— 7 1 9 9V my ” 197 : i be the d rection cosines of the normal a Th e operator a V a p p lie d to a scalar function of position V yields the same res ul t as the direct pr o duct of a and t h e vecto r V V if n , , ” 2a n . . Fo r ( i a o V) V = a ( V V) ( 3 0) . yy th s re a son either operation may be denoted simpl b a VV wi t hout parentheses and no amb igui ty can result from the o m ission Th e two different forms ( a V ) V and a ( V V) may however b e interpreted in an important theorem t h e dire c tio n a Ld e riva t i va e f V in t h e di rect io n a 0 n th a o t h e r ha nd a ( V V) is t ha n o mp o ne nt a f V V_in t h e di re c t io n a Hence : Th e directional derivative of V in any direction is equal to the component of the derivative V V in that direction If V de n o t e gravitational potential the t h e orem b ecomes : Th e directional derivative of t h e po t ential in any direc ti on gives the component of the force p e r uni t mass in that direction In case V b e electri c or magnet ic o tentia l a difi e re n c e of si n must b e observed g . . - . ~ - . . . . p . . VE CTOR 1 50 This m a y NAL YSIS A be writt e n in the form ( a . 7 1 9 1 V) V 9 972 . 9 3 9 17 3 . 9 s k ( 31 ) . s ’ Hence ( a V) V is the directional derivative of the vector function V in the irection a It is possib le to write d o . ( wi thout parentheses V) V a Fo r . a . V V a y the mean ing of -the vector s mbol vs de fi n ed Hence from the present sta ndpoint the e xpression V V can have but the one in terpretation given to it b a y ; V) V (a . Although the operation V V has not been de fi n ed and 1 cannot be at present two formal combin ations of the vector O perator V and a vector functi on V ma be treated These are the ( formal) scalar product and the ( formal ) vecto r p od u c t of V in to V The a re y , y . 9 VV 9 o V is re a 9 d de l do t V; Th e di fferenti a to m 9 9 x by the dot and the cros s . i 9 V x V, de l 9 9 3/ 9 a That z 9 9y a nd 9 9 3 56 z , z c ro ss V XV= I X 9V a: V . b e m g s c a l a r o p e ra t o ra p a ss is 3 9m r 9 9 3/ x i v V . 9 y 9V 9 y 2 9 + k x z 9V 9 z DIFFERENTIA L L UL US OF VE C TOR S 9 73 9 9 x 9 ar 9a ae 9V 9 VI 9 V2 a 2 7 1 z 1 . a 9 n z 9 x av 9 H 9 x J+ . k, x , W , k’ 9v 7 1 9 3 9 k‘ z e nc e More o ve r H VxV en ce 1 9 9 z 9 V2 Th is m a y be ri i 1 51 CA C 9 x i x 9 Vl 9 3; w tte n n the form of a determ nan t y 1 52 VE C TOR A NAL YSIS r is to be unde stood that the operators are to be applied to t h e func t ions V1 V2 V8 when e xpanding the determin ant From some standpoints o b jections may b e brought fo rward a gains t treating V as a sym b olic vector and intr oducing V v and V x V respectively as t h e symbolic s ga g r a n d ve gtg; These objections ma be avoided by L OL M Q LQQM simply laying down the de fi nition that the s mbols V and V whi ch may be looked upon as entirel new operators quite d s tinct from V shall be It , , x y y y i , . , VV— x e 9V gx V x V= i > < B < + k > for practical purposes and for remembe ring fo seems by all means advisable to regard ut e z i a 9 r ul i t m a a 96 9g i l a symbolic vector differentiator Th s symbo obeys the same laws as a vector just in so far as the difi e re n t ia t o rs as . 9 9 w 9 y y q obe the same laws as ordinar scalar uantities 7 3 97 That the two functions V V and V V ha ve ver important physical meanings in connection with the vector fun ction V may be easily recogni z ed B the straight y forward proof indicated in Art 6 3 it wa s seen that the , , . . x y . S VE C TOR A NA L Y IS 1 54 9V m dx d y dz V to tal flu x outward from the cube through thes e faces is therefore t h e algebraic sum of these quan t ties is simpl i Th e y 9 VI 9 V 9 a: d a: d g d z . . i like man ner the flux es through t h e other pa rs of the cube are gV v d x d g d z and k — d cc d g d z In o g e 3! Th e to tal flu x 9 i fa c e s of . z out from the cube s therefore 9V 9V 9 9 g cc q This d cc d g dz . y fl is the net uantit of uid whi ch leaves the cube per unit time Th e quotient of t his by the volume d x dg d z of the cub e g ves the rate of di minution of de n sit This is . i V v= i ° y av av o v s _ __ o 9 9 g 3: 9 W 7 1 9 1 9 z x . e V, , 9 9g i ' z y ecau se V V thus represents t h e di minut on of densit or the rate at which matter is leav ng a point per u n it volume per unit time it is called the dive rge n c e M ax well emplo ed t h e term c o n ve rge n c e t o denote the rate at which uid a p This is the p ro a c h e s a point per unit volume per unit time negative of the d vergence In case the uid s in c o mp res s it le as much matter must leave the cube as enters it Th e total change of contents must therefore be z ero Fo r this reason the characte ristic differential equation wh ch an in c o m p re s s ib e fluid must sat sf is B i , fl . i fl i . . , . l i iy V V 0 0 . y y TH E DIFFERENTIAL i C AL CUL US OF VE C TOR S fl 1 55 where V s the flu x of the uid This equation is often kn own as the h ydro dyn a m ic e qu a tio n It is satis fi ed b any ow of wate r since water is practical l incompressible Th e great importance of the equation for work in e l e c t ric ity is due to the fact that according to Max well s h pothesis electric d s placement obe s the same laws as an incompressible ui If then D be the electric d splacement fl . , y i i fl d . , the operator V x Max well gave t h e na me nomenclat ure has be come widel accepted To Thi s . y ’ y y . y V x V= cu e c u rl . . . i i y i curl of a vector fu nct ion V s t self a vecto r funct on of pos ition in spa ce As the nam e in diga tgs it is closel connecte d with the an gular velocit or spin of the flu x at each point B u t the interpretat on of the curl is neither s o easily obtained nor so simple as that of the di vergence Consider as before that V represents the flu x of a u d Take at a de fi ni te insta nt an in fi nitesimal sphere about an point ( a g z ) At the ne xt instant what has b ecome of the In the fi rst place i t may have moved o f f as a whole Sphere ? in a certain di rection b an amount d r In other words it ma have a translational veloc ity of d r/d t In addi tion to this it ma have un dergone such a deformation that it is n o onger a sphere It may have been subjected to a s tra in b virtue of which it becomes slightly ellipsoidal in shape Finally it ma have been rotated as a whole about some ax is through an angle d w That is to say it ma ha ve an angular velocit the magni tude of whi ch is d w/d t An infi nitesimal sphere therefore ma have an one of three di sti nct t pes of moti on or all of them combined Er}r a translation with efi nite velocit Se c o ri i a strai n w th thr e e i ij de fi nite rates of e ongation a ong the ax es of an elli s oid Th e . i . L g fli . , y l , y . . y . y . y . y . y y y d y l y , . y l y . . . , i , p . VE C TOR A NAL YSIS 1 56 Tha d, i i an angular velocity about a de fi nite axis It s th s third t p e of motion whi ch is given b the c u rl In fac t t h e c u rl o f the flu x V is a vector whi ch has at each point of space the direction of the instantaneous ax is of rotation at t ha t point and a magni tude equal to twice the instantaneous angu lar velocit about that ax is Th e analytic discussion of the motion of a uid presents more diffi culties than it is necessar to introduce in treating the curl Th e motion of a rigid body is su ffi cientl complex to give an adequate idea of the operation It wa s seen ( Art 5 1 ) that the velocity of the pa ticles of a igid bo at an instant s given by the formula y y y . fl . y y dy . r ” i v v(, . x a vo a a i + 1 a r y . r . < + V > 2 j+ a 8 ( a ) x r z k . k + x x x , . e xpand V ( a r ) formally as if it were the vector tri ple product of V a and r Then V r a — v V V a) r V ( ) ( , , . o is a con stant vector v0 V As a Hence the term . 9m ' r 55 - lV x v0 9 r Hmm Therefo “ 1 re in 9 75 + , i van shes . 9y 9z + + é7l 5 is a constant vector it may be place d upon of the differential operator " V . V a a V th e other side . 9 a 2 7j 9 =3 v a — a = 2a r q l i r case of the motion -o f a igid body the curl of the line a r vel o c it at an point is e ua to tw ce the Ia n gu la r velocity in magn tude and n di ec t io n th e y i y i . VE C TOR 1 58 A NAL YSIS ( V ie ) a v Vu Vu x v v = n) x V ( v . ll is to be applied to more th an the one te rm whi ch fo ows it the terms to which it is appli ed are enclosed in a paren thesis as upon the left-ha nd side of the above e uat ons Th e pr o ofs of the formul a ma be given most natu all by e x an di n g the expressions in terms of three assumed unit vecto s 1 k Th e sign 2 of summation will be found con B y mean s of it the opera t ors V v A ve n ie n t tak e th e fo m IfV q i , p r j , r , d summation e xten s over cc demonst rate V x ( u ) v = , g, z . r y x . , . Th e To y . , . V x z i x v a v 9 11. 9; Xv + + 2 2 u 1 1 x + u it 9v > < — 9w Hence To demonstrate V (n o v) v Vu + u V ( X ) V . THE DIFFERENTIAL L UL US VE C TOR S CA C o 1 In k ke m a n n e r Henc v ) ( V e E u = 9 u = — - a v + v r v + u < v > v . r z x u Vv z Vn x( + (V u V x u ) + x v) u > < + r V X V) ( l v . . other formul a a e demonstrated in a simi ar manne 1 4 Th not a tion e 7 ] Th e V ill 0 1 ) V . ( 44) n yi r w be us ed to denote that in appl ng th e Op e ra t o r V to the product (11 v) the quantity u is to b e egarded as c on s ta n t That is the operation V is car ri ed out o nl y p a rtia l ly upon the product ( u In general if V is to be carried out v) pa rtially upon an numbe r of functi ons which occur afte r it in a a enthesis those functions which are con stant for the different ati ons are writte n a fter the parenthes is as subscripts , r . , y pr i , . m ide a m km m m m m p a rt ia l V s o t o s pe a m a y be a vo ide d by e a n s o f t h e fo r u la 4 1 B u t a c e rt a i n a o u n t o f c o p a c t n e s s a n d s i pli c i t y is lo s t t h e re by Th e id e a o fV (n v) is s ure ly n o o re c o pli c a te d t h a n u V v o r 1 Th is . a n d n o ta X (V X u o fa . , . v ti o n ) . VECTOR 1 60 V NAL YSIS A (u i This v) 4 5 ( ) v. formula correspon ds to the follow n g one tion of difle re n t ia l s in the nota y form ul a ( 35) given above ( Art 7 3 ) ma be written in the following manner as is obvious from analog with the correspondin g formula in di fferential s Th e . , V V > < V ( 11 4 y 1 62 VE C TOR ANAL YSIS l form a l y a s if V , 11 n, r l r v we e all rea vecto s x ( x v) = V V r . Then V v : — u z . d i second term is capable of interpretation as it sta n s Th e operator V ha s noth ng Th e fi rst term however is not upon which to operate It the efo re must be trans posed so that it shall have n v as an operand B u t u being ou t side of the parenthesis in u x ( V v) is cons tant for the d fferen Hence ti a t io ns Th e . , , . r x o . i . . u o vV =V ( u o ) v u u V v be a unit vector say a the formul a v and — Ifu , , a o =V ( . / . v a 4 ( 7) a presses the fact that the directional derivative a V v of a vector f u nction v in the direction a is equal to the derivative of the projection of the vector v in that d ection p us th e vector product of the curl of v int o the direction a Consider the values of v at two neighboring po ints ex ir l . . v B ut y 1 i + v2 j+ v3 d v1 = d r o 1 d v2 = d r o 2 d v8 = d r o 3 dv=dr Hence By v k d v = d v1 i + d v2 j+ d v3 k b Hence = dv = dro dv=V (dr . . . THE DIFFERENTIAL CAL C UL US OF VE C TOR S if v denote the value of v at the point ( a value at a neighboring point Or o v = vo + V d ( g, z ) the a nd v ( 49) r is e pression of v in ter Th , 1 63 ms of its value v at a given po int t h e de ls and the di splacement d r is analogous t o the e xp a n sion of a scalar functor of one variable by Ta lor s theorem x y , Th e f CC) i +1 ” ( ) 500 dx derivat ve of ( r v) when v is cons tant is e Th a t is V V (r v) v Fo r ( v v~ = Vr 5c v 1 V q ual to v . 9 7 2 5 1, i + y j+ i + V x e , Vr 0 r ’ r 9 Henc , 0 e v r ( r v) z 2 j+ k, v 3 k = v, = 0 . v v . In like manner if in stead of the fi nite vector r an infi nitesimal vector d r be substi t ute d the result s t ill is , , V (d r By ( 47) v = vo + V (d r V (d r v) 1 Th c 1 is gives another form of ( onvenien t It - v . 1 which is sometime s is also sl ightly more s mmetric a l 4 9) y . m ore VE C TOR ANAL YSIS 1 64 Cons ider L et a mo ving fluid velocity of the uid a t the point (x round a poin t (T g z ) with a sma fl o, At p . ll , o, o r r d d i c phere S r i the inc ement of t me moved the distance In ( 8t r d v, . y V v z , the t ) be t . Su r . is . i p r ll the points of t h s s he e wi St ) vo + d r o i point at the center w ll ha ve move Th e g , 2 each oint of th s sphere the velocit v ( x, ) at t h e time z g, v v , 8t . d dista ce th e n . distance between the center and the points th at were upon the sphere of radius d r at the comm encement of the interval 8 t has become at the end of that interval 8 t Th e d r = d r + d rVv8t ’ locus of the e xtremity of e m nate d r from the e uati ons To fi n d t h e li i q c . Th e dr ’ i it s necessar y to = dr dr o . y Th e fi rst Art 2 . y equation ma be solved for d r b t h e method of 4 7 page 9 0 a n d the solution substitute d into t h e second result w show that the infi nitesimal sphe e , ill r , d rdr . 2 y has been transformed into an ellipsoid b the motion of the fluid during the time 8 t A more de fi nite account of the change that has taken plac e m a y be obta in ed b making use of equation 50 ( ) . y VECTOR ANAL YSIS 1 66 to them s e c tive l y parallel p b ecomes sim m y Then . i the e xpress on a bo ve i point whose coord nates referred to the c ente r of the infi n itesima sphere are Th e l d oc , dz d g, is therefore endowed with thi s velocity will have moved to a new position i t me In t h e . , Th e 8t — dz dy + 5 li y 9 71 2 9y 8t it 1 + dz , 8 t tota t of the points upon the sphere dr li y dg 2 dz 2 goes over into the tota t of points upon the e which the equation is to 3 z 9 " e x llipsoi d of z v3 55 8 statements made before ( Art 7 2) concerning the three types of motion which an in fi n itesimal sphere of fluid ma possess have therefore now b een demonstrated Th e symbolic operator V may be applied several times in succession This will correspond in a general wa y to forming derivatives of an order higher than the fi rst Th e e xpressions found b y thus repeating V w ll all be in de p e n d e n t of t h e a x es because V itself is There are s ix of these dels of the second order Le t V ( re z ) be a scalar f u nction of position in space Th e derivative V V is a vector f u nction and hence h a s a cu l and a divergence Therefore Th e y . . . i . . . r , . V ° V V, V x VV . THE DIFFERENTIA L AL CUL US OF VE C TOR S 1 67 C y are the two derivatives of the second order whi ch ma be obtained from V . V x V V= Th e V e cu 5 2 ( ) . second e xpression V x V Vva n is h e s ide n tic a lly the de ri va tive fa ny o sc a la r f V p o ss e ss e s tio n x unc That is , . n o c u rl j rl This . may be seen by e x pandin g V V V in terms of i k All the terms cancel out ater ( Art 8 3) it will be shown con ve sel that if a vector functi on W poss esses no cu i e f r y L . , . . . , W is the derivative of some scalar functi on V Th e fi rst e x pression V V V when e xpande . o i, j k be c o m e s , 9 V 2 V o VV 9 x 9 V V Sym b o h c a lly , 2 + 2 x 2 + 9y 9 2 9 y L . 9 V i i 9 x 9 V 2 2 + 2 9 y 2 r te ms of + . + 9 z 9 2 9 z 2 2 q ’ i 9 V 2 2 in . 2 2 o 9 9 V d operator V V is therefore the well-known o ap ac e aplace s E uation Th e L l , p rator of e 9 V 2 + y 9 z 2 bec omes n the notat on here emplo ed V VV i When applied to a scalar funct on 0 y d . operator V V iel s a scalar function which is more over the divergence of t h e derivative et T be the te mperat u e in a bod et c be t h e con a n d k the s e c fi c heat Th e du c t ivity p the densit V th e , , L r . , y y L p i . , flo w f is f= — c VT . . VECTOR ANAL YSIS 1 68 i i rate at which heat is leav n g a point per un t volume unit time is V f Th e increment of temperature s Th e . V fdt o d T c k p dt is Let Th is Four . VT . ier s e quation for the rate of ’ ch a nge of tempera a vector function and VI V2 V3 its three com Th e operato r V V of aplace ma be applied to V V be ponente -V i r e p . L , , L If a y , . vector functi on V satis fi es aplace s Equati on each of O ther de ls of the second its three scalar comp onents does order may be obtained by considering the divergence a n d curl of V. Th e di vergence V V has a derivative ’ , . Th e cur l x V V VV = V div V V ha s V V V O f these in i turn a d vergence and a cur v V x V expressions . , x x . cu e curl cu V x , V x V x V V = di v V V l V rl 5 6 ( ) 5 ( 7) V . ide n tic a lly That is , Thi s ma b e is z e ro va n is h e s . y the diverge n c e of th e c u rl of a n y ve c to r seen b y e xpanding V V V in terms of i k ate r ( Art 8 3 ) it will b e shown conversely that f the d vergence of a vec tor function W vanishes identicall y i e f x j L i i i . , . , V W is the curl of some vecto r function V . . , . . VE C TOR 1 70 A NAL YSIS i i V V u is nte re st ng geometric interpret a tion of Th e It depends upon a geometric interpretation of the second derivative of a scala r function u of the one scalar variable cc et u , be the value of u at the point as , et it be required to fi n d the second derivative of u with respect to x at the point et re , and a , be two points e uid s tant from :c Th at is let L L q i . L , = “2 “1 a . . o . o “a 2 ? I “ r is the ratio of t h e difference between the ave age of u at the point s x, and and the value of u at d to the s u a e of t h e distance of the points xl a , from ec Th at o , o i d oc 2 y q r . —O is e as il proved by Taylor s theorem et u be a scalar function of positi on i n space Choose k and eva uate the three mutually orthogonal lines 1 e xpression s L ’ . , 9 2 9 u 2 , Let mo ; (c o ; 9 x j a a l . , z u , 2 di a 1 and be two points on the l ne at a stance a from i , , x and a at the same di stance a from 4 s two points on cc an a s two oints on k at the same d stance a from w a , d , , p j i “1 + u 2 _ u o e . TH E DIFFERENTIAL 9 1 2 9 9 2 9 2 2 7“ Lm 2 a i u x 2 L UL US OF VE C TOR S CA C + 9 2 O VVu a 2 9y 1 71 + 6 a r p s ri and V a e inde endent of the pa t c ul ar ax es chosen this expression m a y be evaluate for a different set of ax es then for st a difle re n t one etc B y add ng together all thes e results As V d ill , u V L et -V 1 + i . u 2 + , o , 6 n t e rm s u u o u become infi nite and at the same time let the sets of axes point in ever direction ss uing from frac tion u u 6 n te rms + 1 2 y n i difle re n t $0. Th e then approaches the a ve ra ge va lue of it upon the surfa ce of a D enote th s sphere o frad us a s u ound ng the point by n i a rr i i . v Vu . LIM -0 a . a 2 y is equ a l to s ix ti mes the limit approached b the ratio of the e xc es s of u on the surface of a sphere above t h e value at the cente r to the square of the ra di us of the sphere Th e same reas oning held in case n is a ve c to r function VV u ( e x cep t for a If u be the tempe rature of a bo co n stant factor which depends upon the mate ri al of t h e V Vu . dy . VE C TOR 1 72 A NAL YSIS q al to the rate of increase of t mperature ( - is p sitive the average temperature upon a small sphere greater than tem perat re at the center center of the sphere is growi g warmer the case of a flow the temperatu e at the center m st remain constant vi dently therefore the condi tion for a steady body ) is e Vu If V Art e u o th e is u Th e . n s te a . dy In . u r E . flo w is V That O V ‘ tt o z . L l q i q y y lr l ddi i i is t h e temperature is a solut ion of ap a ce s E uat on V V it whether M axwell gave the name c on c e n t ra tio n to Conse uentl V V u ma u b e a scalar or vector function be called the disp e rs io n of the function i t whether it b e sca a or vector Th e dis persion is proportional to the excess Of the average value of the function on an in fi nite sima surface above the va ue at the center In case it is a vecto r f u nction the average is a vector ave age Th e a t ons n it a re vector add tions ’ . , . . l r . i . . SUMM ARY i If a OF CHAPTE R III i i vector r s a function of a scalar t the der vat ve of r w th respect to t is a vector uantity who se direction is that of the tangent to the cu rve described b y the terminus of and whose magnit ude is e ual to the rate of advance of that terminus along the curve p e r un it change of t Th e derivatives of the com onents of a vecto are the components of the derivat ve s i r q q i p r . r A . r lr y combination of vecto s or of vecto s and sca a s ma b e dif ferentiated just as in ordinar scalar anal s s e xcept tha t th e difle re n tia t io n s must be performed i n si tu y yi ' . VE CTOR 1 74 A NAL YSIS i denote the position of a mov n g v the velocit A the acceleration If r y rtic l e a p , t the tim e , , , p whi ch one is parallel t o the ta ngent and depen ds u on the v of the parti cle in its rate of change of the scalar velocit path and of which the other is perpendi cular to the tan gent and depends upon the velocit of the a t c e a n d the cur va t ure of the path y y , p ri l . A= s t + Applicat ions v 2 0 . i l to the hodograph in part cu ar motion in a circle parabola or under a central acceleration Application to the proof of the theorem that the moti on of a rig d body one point of which is fi xed s an ins tantaneous rotation abo ut axis through the fi x ed point Inte gration with respect to a scalar is merel the inve se of difle re n tia t io n Application to fi n di ng the paths due to given accelerati ons Th e ope ator V applied to a scalar fun ction of pos t on in Space gives a vector whose rection s that of m o st rapid in crease of that function and whose magni tude is e ual to the rate of tha t inc eas e per un t chan ge of po sit on in tha t direction , , , i . i y . r . r . di r i 9 v=i i 90 a 9 a: i 9 g 5 A 1 9 3/ 9 ’ , z ii q THE DIFFERENTIA L C AL C UL US OF VE C TOR S ope rator V is invari ant of the ax es de fi ned b the e uation Th e y j i, q k , ma It . 1 75 y be ( 24 ) VV d o r d ( 2 5y . ion of the derivative by two me ho ds depend ing upon e q uat ons ( ) and llu tration of the currence of in mathe at cal physics may be looked upon as a ctitious vect r a vector di erentiator obe y s the formal laws of vect rs just in so far the scalar erentiato rs of / bey the fo rmal la s of scalar q uantities Computat VV t 21 i V I m i s oc . V fi f f o It . o di f f as , 9 9 9 cc , 9 y, 9 9 z O w 9V 9V be a un it vector a n the rect on a Ifa i di i 9V i i is the d re ctional deri va t ve of VV . a c V V= If V is i ( a o V) V = a ( V V) . l a vector funct on a V V is the di rectiona o f th a t vector f u nct on in the d rection a i i . av o 9 9 g 0: 9v 9 2: 9 Vl 9 v =i x 9 g 9 V2 9 y < + k > 9 V3 9 9 ’ z 9 9 V3 V VE C TOR A NA L YSIS 1 76 P roof tha t V V is the o fV. dive rge n c e of V an d x V V, V V = d1 V V, ° V xV= V ( + u cu V ( n o ) -u ( u v) o . v V V e ) v =V u +V -v + u -( u V x (u x v) v, ) q — v V o — u — u l d x( 11 u v ) v , 1 4 ( ) 4 2 ( ) , v ( 43 ) Vv + uV . i of the partia de l V (u in wh ch the dif a e performe upon the h othesis that u is fe re n t ia tio ns r x( . u If a be - v Introducti on constan t o (q =v Vu x v) V v + V v, o i = V ) v r yp , ( u i ) v u -u o . a un t vecto the direct onal derivative a V v = V (a 4 7 ( ) i expan sion of any vector funct on v in the neighborho o of a oint ( x y z ) at wh ch it takes on the va ue of v s Th e p 0, o, i o l 0 i 49 ( ) x dr . Appl ication to hydrod n mics Th e de ls y a . of the second order are s ix in numbe r . 50 ( ) d VECTO R ANAL YSIS 1 78 r Obtain p the accelerations of a moving pa ti cle ara llel and perpendicular to the tangent to t h e path and re duce the resul t s to the usual form 3 . y . d a s stem of polar coor inates in Spac e where r is the di sta nce of a point from the origin (1) the m e ridia n a l angle and 0the polar angle obtain the e x pressions for the components of the velocity and acceleration along the radius vector a meridi an and a parallel of latitude Reduce these e xpressions to the ordinar form in terms of x y z 4 If . 0 be r, , , , 5 Show . operator V y , , . , the direct method suggested in is independent of the axes , . that the Art 63 . . i the second method g ven for computing V fi n d the derivative V of a triple product [a b 0] each te m of which is a function of x y z in case a = (r r) r b = (r a) e c = r x f where d e f are constant vectors 6 By . r , , o , , , ‘ , 7 , . Compute V V t . Vis en r z 1 1 r 2 7 , and V x V w h e n V is equal to r and when V is e ua to 5 and show that n these £ cases the formula ( 58) holds 8 9 Compute V V V, V V o . Expand V . x q l o V, . V V a nd V o i , V x V in te rm s show that they van ish ( Art 1 0 Show by e xpandin g in terms of i . . , V> <V > <V = V V 11 . Prove A V (V o W) o V— V j , i, , k j and k th at VV = V A V W+ o of . WA -V V . and (W W) — V (v W) , H C A P TER IV TH E IN TEG R AL CALCULU S OF VECTORS L et W( ) be a vector function of position L et be any curve in space and the rad us Vector ac, y , in z i space r C dra wn from some fi xed origin to the points of the curve Divide the curve into in fi nitesimal elemen t s d r From t h e sum of the scalar product of these elements d r and the value o fthe function W at some point of the element . , . . thus 2 W dr . i r i limit of th s sum when the elemen t s d be come n fi nite in number each approac hi ng z ero is called the lin e in tegra l of W along the curve C and s wri tte n Th e i , , i d x + jd y + k d z , f w 1 ( ) . e de nition of the line integral therefore coincides ith the de nition usual y given is however necessary spec fy which direction the radius vector r is supposed to describe the curve during the integration the elements ave opposite igns when the curve is described in opp Th e w fi fi i l . It to in . dr h s Fo r o 1 80 SS VE C TOR A NAL Y I r site di ection s If one method of descripti on 0 C and t h e other b y . be d y denote b , W -d r . r case the curve C is a closed curve bo undi ng a po t i on of surface the curve will alwa s be regarded as described in such a direct on that the enclosed area appears pos i t ive In y i ( Art y . If f denote t h e y force which ma be supposed to var from point to point along the curve 0 the work done b the force when its point of application is moved from the initial po nt is the line inte g a r of the curve 0 to it s fi nal point y , r , f dr f- rl J r fd : i xo r . ' 0 r line integral of the de ivative V V of a scalar function V (z y z ) along any curve from the point r to the point r is equal to the di f ference between the values of the function V ( w y z ) at the point r and at the point t That is Th e o re m : Th e , , 0 , , o . , VV dr By V (r) z V ( r y, z ) , V ( 900, yo , z o ) . dr V V= d V de fi niti on V = V G) The o re m : V ( ro V Oo ) ' 17 z . 6 0 y ) V( xo r yo , z o ) ‘ ( 2) line integral of the derivative V V of a single valued scalar function of position V taken around a c lo s e d cu ve vanishes r Th e . VE C TOR ANA L YSIS 1 82 value of the integral is therefore a scalar fun ction of positi on of t h e point r whos e coordinates are x y z Th e th e o a together ut . W . dr Let the integr l be taken between two points B , , d V ( x, y , z ) dr it ly near in fi n e . V V dr = dV by defi n ition o W Hence VV . theorem is therefore demonstrated et f b e the force whi ch ac t s upon a unit mass near the surface of the earth under th e influence of gravity et a system of axes i k be chosen so that k s vertical Then Th e L . , j i , — f . L . k g . li i work done by the force when its point of app cat on moves from the position r to the position r is Th e o r r r H enc e i Th e 1? w= force f s said to be derivable from a fo rc e -flen e tio n V when there e xists a scalar func t ion of posit on V such that the force is equal at each point of the derivative V V Evidently if V is one force -function another may be obtained by addi ng to V any arbitrar constant In the abov e e x ample the force -function is i . y more simply Th e fo c e is Or r V f VV , . gz . TH E INTEGRAL C y i AL CUL US i OF VECTOR S 1 83 necessar and suffi cient cond ti on that a force-fun cti on V ( T y z ) e xist is that the work done by the force when its point of applicat on moves around a closed circuit be z ero Th e work done by the force is Th e , , , . 20 i i : l i r y th s ntegra van shes when taken a ound ever clos e contour f= V V = V w If r y inten vanishes er o y by a consta t An d th e . conve sel if di f f nl n VV f Th e . d force-functi on and the work done . const V= w . case there is fric tio n no force-function can e xist Fo r t h e work done by fri ction when a particle is moved around in a closed c rcuit is never z ero Th e force of attraction e x erted by a fi xed mass M upon a uni t mass is di rected to ward the fi x ed mass and is propor nve se square of the distance between the t i o n a l to t h e mass es In . i . i r . i Th s i is the law of universal gravita t on as stated b y N ewt o n It is eas to see that this force is derivable from a force function V Cho ose the orig n of co ordinate s at the center of the attractin g mass M Then the work done s y . i . i . w: r r w=— oM o dr z r — dr d r, 0 M . 1 84 VE C TOR r A NA L YSIS y d a p oper choice of units the constant 0 ma be ma e equal to unity Th e force -function V ma the efo e be chosen as By y r r . di there h a d been several attracti ng bo es the force -fun ction would have been If M 1 , M 2, M 3 , d i where r 1 r 2 r 3 are the distances of the attra cte un t mass from the attracting masses M 1 M2 M 3 Th e law of the conservat ion of mechanical energ requires that t h e work done b y the forces when a point is moved around a cl o sed cu rve shall be z ero This is on the a s s u m p tion that none of the mechanical energ has been converted into other forms of energy during the motion Th e law of conservation of energy therefore requires the fo ces to be derivable from a force-function Conversely if a force function e x ists the work done by the forces when a point is carried around a closed curve is z ero and conse uently there is no loss of energy A mechanical system for which a force function e xists is called a c o n s e rva tive system From the e xample just cited ab ove it is clear that bodies mov ng under the la w of universal gravitation form a conservative s stem at leas t so long as they do n o t collide et W ( as y z ) be any vec t or function of position in sp ace et S b e any surface Divide this surface into in fi n it e s im a l elements These elements may be regarded as plane and may be represented b y in fi n itesimal vecto s of whi ch t h e di rection is at each point the direction of the normal to the surface at that point and of which the magn i tude is equal to the magnitude of the area of the in fi nitesimal , , , , y , y . r . . q . . . L L i y . , , . . r VECTOR A NAL YSIS 1 86 gives the amount of that substance which is passing through the surface per unit t ime It was seen before ( Art 7 1 ) that the rate at wh ich matter was leaving a point p e r unit volume per unit time was V f Th e total amount of mat ter which leaves a closed space bounded b a surfac e S er unit time is the ord nar triple inte gra l . . y . i y V p fd v f o . y Hence the ver important relation conn ecting a surfac e in tegral o f a flu x taken over a c lo s e d surface and the volum e integra of the divergence of the flu x taken over the spa c e encl osed b the surface l ' y V fd v . i 7 a f a r i y l Wr tten out in the notation of the ord nar c a c ul u s this becomes X d d z y [ d or e x . dx d 9 17 + o z 9 3, oz dy ] dx d y d z ( 8) where X Y Z are the three components of the flu x f Th e theorem is perhaps still more familiar when each of the th e e components s treated separatel , . , i y ff d This dy = r . fff d x d y dz . is known as G a us s s Th e o re m It states that the surfac e integral ( taken over a closed surface ) of the product of a f u nction X and the cosine of the angle wh ch the e x te rior normal to that surface makes with the X-axis is e u a l to the volume integral of the pa tia derivative of that funct on ’ . i r l q i THE INTEGRAL CA L C UL US OF VE C TOR S 1 87 y l with respect to x taken throughout the volume enclosed b that surface If the surfa ce S be the surface boun di ng an nfi n tesima sphere or cube i i . f da = V o W here d o is o fd v l the vo ume of that sphere or cube 1 fda do H ence . . is e quation may be taken as a fi of the vergence f divergence of a vector function is equal to the limit approached by the surface integral of f taken over a sur Th V de . di n i tio n Th e f face boun ding a n in fi nitesimal body divided by that volum e when the volume approaches z er o as its limit That is . f a 0 a ( 1 0) . y From thi s de fi nition which is evidentl indepe ndent of the a x es a ll the properties of the divergence may be deduced In order to make use of this de fi nition it is necessary to develop at least the elements of the integral calculus of vecto s before the differentiati ng operato s can be tr eated This de fi nition of V f consequently is interesting more from a theoretical than from a practical standpoint Th e o re m Th e surface integral of the curl of a vector f u nct ion is equal to the line integral of that vector function taken around the closed curve b oundi ng that surface . r r . . . x W-d a W e a r. 1 ( 1) is is the celebrated theorem of account of great importance in all branches of mathematical phys ics n mber of di erent pr ofs will be given Th u Sto ke s f f o . . On it s 9. ' VECTOR A N ALYSIS 1 88 First Proof : a small tri angle 1 2 3 upon the su rface et the value of W at the verte x 1 be W Chap III the value at an neighboring point is F i ( g S y . Then b Consider L . . y , 0. w Sr whe re the s y di m bol 8 r has been introduced for the sake of s t i n guis hin g it from at r which is t o be used as the element of inte gration Th e inte gral of W taken aroun d the 1 9 3 is . FIG 32 . l . 1 x 8 r dr . A Th e fi rs t t e rm f w -d r o A i l r i van s hes be cause the integral of d r around a c osed fi gu e n th s case a small tri angle is z ero Th e second te rm i , m L v i y va n sh e s b virtue of ( 3 ) p a ge . . 8 r) 1 80 . . dr Hen ce , Th e x \ VE C TOR A NAL YSIS 1 90 2 second member W 7 8 r da rl is the su fa ce inte n of the cu of W E . da V x W- V x 8 W o da . i r addi ng together the line integrals wh ch occur in the fi s t member it is necessary to notice that all the sides of the ele mentary triangles e xcept those which lie along the boun di ng curve of the surface are traced twice in Op p o s ite di rect ons Hence all the terms in the sum In i . y which arise from th ose sides of the triangles l in g wi thin the surface S cancel out leaving in the sum onl y the terms whi ch arise fr om those sides whi ch make up the boundi ng curve of the surface Hence the sum reduces to the lin e in t e gra l of W along the curve whi ch bounds the surface S , . . Henc W e . da L Second l . L . . Fro . 33 . y Proof : et C be an closed contour drawn upon the surface S F It wil be assumed that C ( ig is continuous and does not cut itself ’ et C be another such contour near to 0 Consider the variation 8 which takes place in the line integral of W in passing from the contour C to th e ’ contour C . NTEGRAL TH E I s f w . f d A L C UL US C W dr = o OF VECTOR S 1 91 dr 01 e w 8 r W ( ) e r) fW- Sd r + = d W8r + f 8 Wdr . d 8r W- 8dr = d 8r . H ence o 8 dr f f W-d 8 r = d( 8 r) W- Sr d W- . r expression at ( W 8 ) is by its form a perfect di fferential Th e value of t h e integral of that ex press ion wi ll therefore be the difference be tween t h e values of W d r at the end and at the beginning of the path of integration In this case t h e integra is taken around the closed contour C ence Th e . l Hence . . f W -Sd r d W-8 r , H VE CTOR 1 92 Subs tituting A NAL YSIS these va lues a w dr i o Br 8r il r sim a terms in y and z B ut y b ( 25) page . 111 CW 3 io d 9r dr i- a r i o dr . x f f Henc e 9 W-d r = S ‘ Z o a e dt s imilar terms in f W S o f dr = y an d z W i ri r it w ll b e seen that d s the element of a c along the curve C and 8 r is the distance from the curve C to the curve C Hence 8 r d r is equal to the area of an ele mentar parallelogram included between C and C upon the surface S That is In Fig . 33 x ’ y . ’ . B r f S Let the curve x W -d r dr = da, s x da W- . il l starting at a point 0 in S ex pan d unt it coincides with the contour boundin g S Th e line integra C . f w-d r y will var from 0 at the point f ol r W- O valu e VE C TOR 1 94 AN AL YSIS defi nition of V >< W which is independent of the axe s k may be o b t ained b y applying St okes s theorem to an in i Pass a plane C onsider a point P fi n ite s im a l plane area t hrough P and draw in it concent ic with P a small circ e of area d a W dr V x W da ( 1 3) A , j ’ , r . , l . , . o o When d a has the same direction as . W the value of the line integral w ll be a max imum for the cosine of the angle Fo r th is between V x W and d a will be e ual to unit value of d a i V x q , y , w Hence the curl d . r of a vector function W has at each point of space the direction of t h e normal to that plane in which the line integral of W t aken about a small circle con centric wit h t he point in ques t ion is a max imum Th e mag n it u de of the curl a t the point is equa l to the magnitude of that line integral of max imum value di vided by the area of the c ircle about whi ch it is taken This de fi nition lik e the one given in Art 8 1 for the divergence is interesti n g more from theoretical than from practical cons ideration s Stokes s theorem or rather its converse ma be used to de uce Maxwell s e quations of the electro-magnetic fi eld in a simple manner et E be the electric force B the magnetic induction H the magnetic force and C t h e flu x of electri cit p e r unit a ea p e r unit time ( i e the current densit ) It is a fact lea rned fro m e xperiment that the tota el e ctro mot ive fo ce a ound a closed c ircuit is e ual to the negati ve of t h e ate of chan ge of to tal magnetic induction through th e ci cuit Th e tota electromot ve force is the l ne ntegra of the e ectr c fo c e ta ken aroun the circuit That is V > <W . . . d y ’ ’ . r r r L . y , , r r l l i r . y , . . q d i l . i i . Edr . l NTEGRAL TH E I AL CUL US C i VE C TOR S OF 1 95 total magnetic induct on through the circuit is the sur face inte gral o f the magnetic induction B taken over a surfa c e bounded b the circuit That is Th e y . B ri Ex pe m ent -d a r the efore shows that Eu h t a r m o H ence by t h e fiL da B r conve se of Stokes s theorem ’ V x E a — l w flE y is a so a fact of e xperiment that t h e work done in carr ing a unit positive magneti c pole aroun d a closed circuit is equal to 4 7: times th e to ta elec t ric flu x through the cir cuit Th e work done in carr ing a unit pole around a circuit is the lin e integra of H aroun d the circuit That is It l l y . . H l o dn y i i to ta flu x of elec tri cit through th e c rcui t s the surfa ce in tegra of 0 taken over a s u face bo u n e d b t h e circ ui t That s Th e . l i r C da ' Ex p rim nt the efore tea hes th t e e c r H o a dr = 4 7r Co da . d y 1 96 con ve B y th e r V ECTOR se A NAL YSIS of Stokes s theorem ’ V X H =4 p C 7r With a pro er inte rpretation of . r c u rent 0 a s t h e dis placement current in addit on to the conduction current an interpretation dependi ng upo n one of Maxwe s pri mar h pothes e s this elation and the preced ng one are the fun da mental equati ons of Maxwell s theor in the form u sed b Heaviside a n d H e rtz Th e theorems of Stokes an G a uss ma be use to strate the ide n tities y , i th e r y ’ VV x W= 0 di v , rl ing to G auss i d =O . . V x Wda ’ . r to Sto kes s theo em Hence Wdr W -d a VV d i d V V= 0 v ’ V x Apply ’ s theorem VV A ccord n g y y , , e cu cu - Accord i y V x V V = O, ’ ll d . , . W ' o o dr . ' th s to an in fi n ite s im a l s p h e re Th e su rfa c e bounding the sphere s cl o sed ence its boun di ng cu ve e uces to a o nt ; a n d t h e integra around t to z e o pi i . H l r rd . i V V X Wd v ' O V V X O r , W o W= 9 . . d r = 0, 1 98 i VE C TOR ANAL YSIS y i y those connected w th integra t ion b parts n o rdinar calculus The are obtained b y inte grating both s des of the formul a page 1 61 for di fi e re n tia t in g y . V ( Hence I p res ion u ex d r ) u v ' V i . , , Th e “ o u dr [ : V u u v ' dr v r ] v u u dr dr o . . 1 4 ( ) r [ s u n ] represent s the difle re n c e between the value of ( u v) at r the end of the path and the va lue at t the beginning of the path Ifthe path be closed ' , o, , u V vo ' dr . v V u o dr . Second v x ( u d a ) v ff = u V - Xv da + Vu - xv da . 8 V u xv da f n O v dr— fq S da, X v- 1 ( 5) THE INTEGRAL CAL C UL US OF VECTOR S V x (q d a ) V v V X u u 1 99 da . Hence V u x V v ' d o u a V vdr vV u ' , v Fourth V d r ( 1 6) . ( o u v) d o u V vd v + V ' u vd v . Hence u vd v = V- ff u v da — S V u vdv ' s u v -d a — S fff V fff u v u o vdv , vdo, v — ff H ence v(V ff v u u x v) d x v da v f — s V u fff — V fff V u -v dv f . ( 1 8) u these formul a which contain a triple integral the surface S is the closed surface b oun di ng the bod thr oughout whi ch the integration is performed Ex amples of integra tion b y parts like those ab ove can be m ul tiplied almost without limit O nly one more will b e given here It is known as G re e n s The or e m and is perhaps the most importa nt of all If u and r; are any t wo scalar func t ions of position In a ll y . . ’ . . , VE C TOR A NAL YSIS 200 v (u V o) = u v V V ( v ) - VV u v= VVv ' , JV V V(v u q O ) a . V ( vV u ) — — c ffq v vu d v; o f Vva d o dv f , . e u o v vd v ' ' fv s « u V , u Vv+ V V e dv ff By u ' ' ° Henc Vv+ =Vu . : u da — q i da ffv u o V u d V V v d o, v . ubtractin g these e ualit es the formula — v v v ( V V u o v ) u ( V u i ( 1 9) ( 20) ff dv ' ' o r— v v ) da u o . obta ined B y expand ng the e xpression in terms of i i k t h e ordi nar form of G reen s theorem ma be obta ined A further gene aliz ati on due to Thoms on ( ord Kelvin) is the following is . y V u r v where w is u i q d a — vo , . ff V fff V v dv= jf y L ’ , da u V 2 1 ( ) v ' a th rd scalar function of po sition Th e element of volume d v has nothin g to do w th the s ca a fu nction e in these equations or in those th at go before Th e us e of o in these two di f ferent senses can ha dl g ve se t o an y misunderstanding In the precedi ng articles the scalar an vector fun c ti ons which h ave been subject to treatment h ave been su i lr r y i ri d p . . . VE C TOR A NAL YSIS 2 02 the value along one path being the negative of the value along t h e other Th e integral around the circle whi ch is a closed curve does not vanish but is e ual to i 2 It might seem therefore the results of Ar t 7 9 were false and that consequently the enti re bottom of the work whi ch follo ws fell out Th s however s not so Th e di fii c u lt y is that the function q . , z z 7 r. . i . i V . ta n : is not single-valued At the point funct on V takes on not o nl the va ue i y . V tan : l fo r i ns tan ce , th e 1 1 but a whole series of values I + k 7r, where la is any positive or negative integer Furthermore at the origin which was included bet ween the t wo semicircul ar paths of integrat ion the function V becomes wholly inde terminate and fails to possess a derivative It will b e seen therefore that the origin is a peculiar or s i n gu la r point of the V If t h e two paths of integration f om function r 1 O) to had not included the origin the values of the integral would not have di ffered In other words the value of the integral around a closed curve which do es n o t n i c lu de th e o ri i n vanishes as it sho ul d g . , , . . , . . vitiates the results obtained let it be considered as marked by an impassable barri er An closed curve C which does y , . NTEGRAL CALCUL US TH E I y OF VE CTOR S 2 03 ma shrink up to no t hin g without a break in its continui t y ; b u t C can only shri nk do wn and fi t closer and closer ab out the origin It canno t b e shrunk down to not hing It mus t always remain encircling the origin Th e c urve C is said to be re du c i ble C i rre du c i ble In case of the func t ion V then it is true that the integral taken around any re du c i ble circuit C vanishes ; b u t the integ ra l around an i rre du c i ble circuit C does n o t vanish Suppose ne x t that V is any function whatsoever et all the points at whi ch V fa ils t o b e con t inuous or to have con t in u o u s fi rst partial derivatives be marked as impassab le Then an circuit C whi ch contains within it no b arriers such point may be shrunk up to nothing and is said to b e re du c i ble ; but a circ ui t whi ch contains one or more such points cannot be so shrunk up wi t hout b reaking it s continuity and it is said to b e i rre du c i ble Th e theorem may then b e stated : Th e li n e in tegra l of th e de riva t ive V V Of a n y fu n c tio n V va n is h e s a r o u n d a n y re du c i ble c i rc u i t C It may or ma not vanish around an irreducib le circuit In case one irreducible circuit C may b e dis tort ed so as to coincide with another irreducible circuit C without passing throug h an of the singular points of V and wi t hout break in g its continuity the two circuits are said to b e re c o n c ila ble and t h e values of t h e line in t egral of V V about them are the same A region such that a n y closed curve C within it may b e shrunk up to nothing without passing through any sin gu ar oint of V and without b reaki ng its cont inuity that is a region every closed curve in which is reducib le is said to be a c y c lic Al l other regions are c yc lic B y means of a simple device any cyclic region may b e ren dered acyclic Consider for instance t h e region ( Fig 3 4 ) e u closed between the surface o f a cylinder and the surface of a cube which contains the cylinder and whose bases coincide with those of the cylinder Such a region is reali z ed in a room . . . , . , , y L . . y . . y . . y , . l p , , , . . . , , . . 2 04 VE C TOR i A NAL YSIS fl wh ch a column reaches from the oor to the ceili ng It is evident that this region is cyclic A circuit which pass es It cannot be contra cted to a ound the column is irreduci b le nothing without breaking its continuit If now a diaphragm be inserted reachin g from the surface of t h e cylinder or column to the surface of the cube the region thus forme bounded b y the surface of the c linder the s urface of the cube and the t wo side s of the O wing to the ins a r dia p h ra gm is ac clic F G 84 ti on of the di aphragm it s no longer poss ible to draw a circ ui t which shall pass completel around the cyl inder — the diaphragm revents it Hence ever cl o sed cir cuit which may be drawn in the region is reducible a n d t h e region is acyclic In like manner a n y region ma be rendered acyclic b inse ting a suffi cient number of di aphragms Th e bounding surfaces of the new region consist of t h e b ounding surfaces of the given cyclic region a n d the two faces of each di aph ragm In acyclic regions or regions rendered acyclic by the fo e going device all the res ults contained in A rts 7 9 e t se g hold tru e Fo r cyclic regions they ma or ma not hol true To enter further into these questions at this point is unnecessary Indeed even as much di scussion as has b een given them already may b e super uou s Fo r the are ques tions which do not concern vector methods an more th an the correspondi ng Cartesian ones They belong properly to the subject of integration itself rather tha n to the particular notation which may be employed in connec t ion with it and which is the primar o b ject of e x position here In th s respect thes e uest ons are s mi ar to uestions of gor in r . . y . d y I y p , , i . . . . y y . y . r y . r . y . ’ y d . . . . fl , y . y . , q i y i l q . ri i . 2 06 Th e VECTOR A NAL YSIS length of positive 7 r and 12 wi ll be assumed to be . Vr ’ is then r1 2 Consider “3 2 ( 31 2 0 (2 2 90 z 1) 2 ' the triple integral 0 1 ( x2 r 2 5 z 1, V05 2 , ) z 2 ) r d x z d l/s d z r i integra t ion is performed w th respect to the variables i x z that is with respect to the b ody of wh ch V y 2 2 2 D uring represents the densit ( Fig ‘ t h e integration the point ( x 1 I l v z l ) re main s fi x ed Th e integral I has a defi ni te value at each de fi nite point (x 1 y 1 z l ) It is a function of that point Th e in F G 35 if t e rp re t a t io n of this integ a I s eas the fu nction V be regarded as the density of matte r n space Th e element of m a ss d m at ( x 2 y2 z z ) is Th e , , y , . , . , , rl i y i . . r . . , V ( n a , y2 , dm r z 2) , . , d wa d yz d z 2 Vdv . integra l I is therefo e the sum of the elements of m a ss in a body each di vided b it s s tance from a fi xed point Th e , ( 33 1, $ 1 1, z 1 ) y di ° This is what is term e d the p o te n tia l at due to the bod whos e ens it is y d y 17 02 5 z , a th e po i t ( n x1, yl , z ) l ) limit of i t grat on in the inte ral l oked at in either of ways the st place the y may be reg ded coincident with the li its of the body of w ch the densi y hi i deed ight seem the most nat ral set limit other h d the integral be Th e s i n e t wo ar In . m t s . . T s On t h e n o fir as V is of g I ma y b e m an hi u I ma y THE INTEGRAL regar ed as taken over AL C UL US OF VECTOR S 2 07 C value of the int gr e al p is t h e same in both cases Fo r when t h e limits are in fi n i te the fun ction V vanishes identi call y at every point ( x 2 y2 z z ) situated outside of the body and hence does not augment the value of the integral a t all It is found most convenient to consider the limits as in fi nite and the inte gra l as e x tended over all space Thi s saves the trouble of writing in special limits for each particular case Th e function V o f i ts elf then practically determ ines the limits owing t o its vanis hing iden ti cally at all points unoccupied b matter Th e O peration of fi nding the pote ntial is of such fre uent oc currence that a s p ecial symb ol P o t s used for it d a ll a ce s Th e . . , , . . . y . q , Pot y z ? y” V r ) p e x , , i . d y, d z z . p s mbol is e ad the otential of V Th e otential P o t V is a functi on not of the variab les x 2 y2 z 2 with re gard t o which the inte grati on is performed but of the point which is fi x ed during the int e grati on hese 22 T z ( 1 yl l ) Th e function V variables enter in the e xpression for and P o t V t herefore have di fferent sets of variab les It may be necessary t o note that although V h a s hitherto be en regarded as the densit of matte r in space such an interp etation for V is entirel to o restri cted for convenience Whenever it becomes necessary to form the inte gral Th e “ ” . , , , , , . , . y y r ” x ” if” Z 2 ) dz l cry , , , . dz 5 , se , my . of any sca a r function V no matter what V represents that inte gral is call ed the potential of V Th e reas on for c a lling such an inte gral the pote ntial even in cases in wh ch it h a s no connection with ph s ical potent a is that it is formed la was the true po tent al and a ccord ng to the same form a , , . i y l il i i VECTOR ANA L YSIS 2 08 y b vi rtue of tha t formation ha s certa in simple rul es of Ope tion which other t pes of in t egrals do not possess Pursu a nt to th s idea the potential of a ve c to r f un ct o n i y 9 , 2, z a a i . W ( 372 r ) q this case the i ntegral is the sum of vector uantiti es and is consequently itself a vecto r Thus the potential of a vector function W is a vector function just as the po tent al of a Scalar function V wa s seen to be a scalar func tion of posi tion in space IfW be reso ve into its thr ee com ponent s In i . , l d . W ( x2 , 9 3” Z 2) i X ( 002 , z a) + jY ( wa ( x2 kZ Pot , W = i P o t X + jPot Y + k Pot Z . z 92, r l , I la ) z ) z z ( 2 4) . r potential of a vector function W is equal to t h e vecto s u m of the potentia s of its three components X Y Z Th e potent al of a scalar function V e x sts a t a o nt r e ‘ z ( l Il v 1 ) when a n d o nl when the nte gra Th e l i , y , i l i , pi . , . Pot V l taken over all spa c e converges to a de fi nite va ue If for instance V were ever where constan t in Space t h e in t e gra l would become greater and greater w t hout m t as the limits of integration were ex tended farther and farther out into spa ce Evidentl therefore if the pote nt a is to e x s t V mus t approach z e o as its m t a s t h e po int (212 y, z g) ecede s inde fi ni te y A fe w mportant s u ffi c ent condi t ons fo r t h e converge nce of the potent al ma b e o bt a n e by t an s forming t o pola c o b rdin a t e s Le t y , r . r r l . r i y i , . li i il li i i . ‘ y i i ' i i d , . VECTOR A NAL YSIS 21 0 f u nc tio n If the V re m a we a kly th a t the p ro du c t re ma a s in s fi wh e n n i te r f fin ite in or i th e n the in tegra l o rigin a re c o n c e rn e d. to th e a r a s regio n s n e a r f so Vr a c h e s z e ro , ro pp a i t be c o me inf in ite c o n ve r es g Fo r let Vr < K =R r r : Vr s rn ddr a s a s < jfd r r a s a s =0 =R Kdr r r a s a s n : KR . =0 Hence the t iple ntegral taken over all space inside a spher of ra us ( where is now supposed to be a small quant ty ) is less than and consequently converges as far as K i r di e R R 4 7r i R regions near to the origin which is the poin t ( sol concerned , z yl , are ) l . If i e . a y p o in t ( x 2 , y 2 , 2 2 ) p o in t ( x 1 , y l , z l ) , th e t a n t he . we a kly th a t th e p ro du c t x ( 2 , y2 , Z 2 ) by th e s q u a (x2 f o re n o t c o in c i de n u n c tio n f t h e va lue f th e O fi V be c o m e s i n fV o dis ta t wi th th e nce a t a p o in f th a t O t o rigin , n i te so nea r to p o in t fro m a in s r e m n i a a h e n i h a i a t e s d a e ro c h e s z e r o t t t s t n c ) f y pp 2 th e in te gr a l c o n ve rge s a s fa r a s re gio n s n e a r to th e p o in t ( x 2 y 2 Z 2 ) , 2, Z , , , proof of this statement is like those given before These three con di t ions for the convergence of the integral P o t V are suffi cient They are b y no means n e c e s sary Th e integral may converge when they do not hold It is however indispensable to know whether or not an integra under discussion converges Unless the tests given ab ove show the conver gence more stringent ones must b e resorted to Such however wi ll not be discussed here They b elong t o t h e theory of integration in gene ral rather than to the a re c o n c e rn e d . Th e . . l . . . , . , , . y THE INTEGRAL CAL CUL US OF VECTOR S 21 1 theor of the in t egrating operator P o t Th e dis cussion of the convergence of the potential of a vector function W re duces at once to that of its th re e components whi ch a re sca ar functions and may be treated as above Th e potential is a function of the variables x1 y l a 1 which a re con stant with respect to the integration et the value of the potential at the point ( x1 y 1 z l ) b e denoted by . l . L , . , P Ot [ i Th e fi rst is m a . , , l 311 , a partial derivat ve of the potentia with respect to therefore a P Ot V 9 33 P O t V [ ]z LIM A 1 x 1 1 O i A value of this li mit ma device ( Fig C onsider the potential at the point Th e 21 , P O t V [ l y r, 3 1 ’ yl ’ A 93 1 y be determ ined by 21 a , ( 2 5) a simple . A 901 , z y ) r i due to a certain bod T Th s is the same as the potent a at the po int ( x 1 , y1 , z l . il : M ) Fro . 36 . y y due to the same body T di splaced in the negative di rection b the amount A ee l Fo r in fi nding the potential at a point P due to a bod T the absolute positions in space of the bod T and the point P are immaterial It is only their positions re l a ti ve to e a c h o th e r which determines the value of the poten tial If both bod and point be tra nslated by the same amount in the same d rection the value of the potent al is un changed B u t now if T be displaced in the negative di rect on b y the amount A ac the v a l ue of V at each point of s a c e s chan ged from 7 7 05 2 yr z z ) to V s A 502 9 2 where A x 2 A re ] y . . y . i i , , . i p i . , , , VE C TOR A NA L YSIS 21 2 Henc e P ot V [ H A ‘n S in J P O t V [ e nc e Aa l i P 015 V( x [ 2 31 ”: A 3 1 9 Vi a 3; A x 2 ’ ga P L O Vs [ , z e ” 21 . h o 31 31 y” A 33 1 O d will be found convenient to intro uce th e li mits o f inte gra t ion et the portion of space originall fi ll ed by t h e body T be denote d by M ; and let the portion fil led b t h e body after its translation in the negative d rection through ’ ’ Th e regions M and M the dista nce A re ] b e denote d by M overlap et the eg on common to both be M ; and let t h e ’ ’ Then em a nder of M be m ; the remainder of M m It L . y y i r i L . r i . , M=M + m Pot A 7 1 6 73 1 ” 2 , V( a 29 2 , A x 29 y 2 9 s r 2 2 ) d v2 12 Axe , Se i z e ) " Pot . 12 V ( x2 , y a , e V ac a) : 2 y” 2 ) d v2 12 V ( 932 ya Hence ( i As in di c e s a 25) 2 ) do V 03 2 9 2 i , g . 2 2 ) g be comes when A x , s replaced b y fo llo wi n po t e n t ia ls be e n dro ppe d ll t h e ha ve 2 a re fo r t h e po in t x1 , y 1, it s z de q g e ual A a i the bra c k n, e t a nd 21 4 Then if it be VECTOR A NA L YSIS ssumed that the region upon the surface b oun di ng T vanishe s T is fi n a LIM é l) A xe — r V( x 2 y2 ’ r 12 ’ e and th a t V d v2 = 0 s 12 it z ? ) d s , ” 2 O . Cons equently the expression for the derivati ve of tial reduces to merely poten Pot d vz Th e p a rti a l de riva tive of the p o te n ti a l th e f o a s c a la r f u n c tio n V l to th e p o te n tia l of the p a rtia l de ri va tive of V Th e de ri va tive V of the p o t e n t ia l of V is e q u a l to th e p o te n tia l of th e de riva tive V V is e qu a . . V is statement follo Pot V Pot V V 2 7 ( ) : immediately from the former As the V upon the left-hand side applies to the set of vari ables xl Z‘l v z l it may b e written V I In like manner t h e V upon the right-hand side ma be written V2 to call atten ti on to the fact that it applies to the variables wz yz z z of V Th , ws . y , . , To i, j i VI Pot V = Pot V2 17 y demonstrate th s identi t k , V . y d ma be e xpa nde in . 9 Pot V 9 I f !1 — i Pot + , jPot +k Pot . j THE IN TEGRAL CA L C UL US y OF VECTOR S 21 5 are consta nt vectors the may be pla ced under the sign of integration and the terms ma be coll e cted Then by means of ( 2 6) As i, , k y . V 1 P 0t V = P Ot V z V The Vx c u rl f it/no t io n W dive rge n c e di ve rge n c e V a nd a re e u a q l r e sp e c h a t u n c t io n t f f o the p o te n tia l ti ve ly to th e p o te n tia l f o f the o vec to r a c u rl a n d . W P ot c u e V l P o t W Pot V2 W di v Pot w P o t d v w curl Pot : : i These f o i y , 9 . relat ons ma be proved in a mann er analogous to the It is even possible to go further and fo rm the de l s a bove o f h gher order i . V 0 V Pot V V Pot V- : Pot V V V, ( 3 0) W = P o t V -V W Pot Wfi Pot V V VVW V x V x P o t W= P o t V xW (31 ) , 32 ( ) , ( 33) . pon the left might have a sub script 1 attached to show that the differentiations a re performed with res pect to the va riables x1 gl z l and for a simil ar reas on the de ls upon the right might have been written wi t h a sub script 2 Th e esul ts of this article may be summed up as follo ws Theorem : Th e dife re n ti a tin g op e ra to r V a n d th e i n tegr a tin g Op e ra t o r Pot a re c o m m u t a ti ve In the foregoing w ork it has been assumed that the egion T wa s fi n ite and t ha t the function V wa s everywhere fi nite and continuous inside of the region T and moreover decreased so as to approach z e ro continuously at t h e surface bounding that region These restrictions are inconvenient Th e de ls U , , , r . . r . 21 6 VE CTOR y r y A NA L YSIS i r ma be e moved by ma k n g use of a su fa c e Th e deriva tive of the po te nti al wa s o bta n e d ( a ge essenti all th e form and i p it l n e g ra . 2 1 3) in 9 Pot V 9 x 1 V032 1 LIM A A xe — O A x2 , ya , 5 2 ) s r 12 VG” ? y 2 ’ i l q l 2 2 )dU ‘ 2 i a d re cted element of t h e surface S bo u nd ng t h e ' egion M Th e e ement of volume d v z n t h e reg on m is there fore e ua to Le t d a b e r . i d v2 = A x 2 i He c V023, n e A 7 VO5 2 A 7 Th e da lmn e e e t l i x 2, ' f lz a z 2 ) 23 2 2 r i i o V( x2 ’ 7 V03 2 » ya , z a ) s ) d v, 32 of vo ume d v2 n th e eg on A Conse . 93 » 2 12 i i da z m i da 0 . is q l e ua to . 2 ) d ?) a 12 day quently 9 Pot V i o dS . 34 ( ) VECTOR 21 8 A NA L YSIS L thi s point with a small Sphere of radi us R et S den ote t h e s u rface of this Sphere and M all the region T not included within the sphere Then . . 9 Pot V 9 x1 . 1 ii i the cond t ons mposed upon By . da . V Vr < K h E de d¢ = 2 da C onsequently 7r K . ll when the sphere of radi us R becomes sma er and smaller the surface integra l may or ma not be come z ero Moreover the volume integral y y y . ma or ma not approach a li mit when R becomes and smaller ence the equation . H sm ll a er 9 Pot V 9 58 y r 1 r i not alwa s a de fi nite meaning at a point of the eg on T at which V becomes in fi nite in such a manner th at t h e product V r e m ain s fi n ite If however V remains fi nite at the point in uesti on so that the product Vr approaches z ero the constant K is z e o and the surface in tegra bec omes sma ller a n d sma er a s R appro a che s z e o Moreove t h e volume inte g a ha s , q . , r l . , r rl 1 ll r THE INTEGRAL OF VECTOR CA L C UL US appro a ches a defi nite li mit as R becomes infi ni tesim sequently the equati on OP o t V Pot 67V 9 x S al 21 9 Co n . 9 wz 1 holds in the neighborhood of all isolated poin ts at which V remains fi nite even though it be dis continuous Suppose that V becomes in fi nite at some single po int z x ( 2 y 2 z ) not coincident with ( 1 1 gl z l ) A ccording to the condi ti ons laid upon V 2 W K . , , , , i . where l s the distance of the point ( x2 near to it Then the surface inte gra l . i d nee not be come z ero and o c y oo z rom a po int f ) z da ly the e quation OV 9 l need not hold for an point f V becomes in fi nite at gz i y2 , consequent 9 PotV A , , Vl gl , z 2 < x z 2 r i of the eg on ) l . B ut in such a mann er th at K , then the surface integral will approac h z ero a s its limit a n d the equat on will hold Fin all y suppose the fu ncti on V remai n s fi nite upon the surfa ce S boundin g the region T but does not van ish there In th is ca se there exists a surface of dis cont nuiti es of VI Within th s su face V is fi nite ; w thout it s z ero Th e i . , i r i i da , i i . . VE CTOR 2 20 doe s not vanish general in 9 A NA L YSIS H ence the e q atio Pot V 9V 9 9 x 1 a c a nnot hold Similar reasonin g ma be applied pa rtial derivatives w th espect to co ] t h e resu ts it is seen that in general y i r . l Vl L et n u . s to , g] , Pot V = Pot V, V i ( 35 ) fun ction n spa c e and let it be gra n t e d th a t Pot V e xists Surround each point of spa ce at whi ch V ceases to be fi ni te by a sma ll sphere et the surface of the sphere be denote d by S Draw in spac e a thos e s u rfac e s which a re surfa ces of di sconti nuit of V Le t thes e su faces also be denote d b S Then the formu a ( 8 5 ) hol s whe e the surfac e integ al s taken over all the s u rfaces which have been e signated b S If the inte n t aken over all these surfaces van shes when the adi of the sphe es a b ove mentioned becom e nfi n i tesi m al then V hs any , L . . . y r i d y i i r This Pot V Pot : n i te su re ly or o d r uc p t ho ld r pp li ed a t fi be c o me s in VI re m a dis c o n t in u it y, as r d l . r i . r , V2 V . fo rmul a Pot V, V: VI P o t V : fi y . VI wil l ll be c o me s a p o in t n it e ins a t ( x 1 , Yb 2 V f i ) 1 p o i n t ( x) , Ya , n d if V o s p se s ses a re m a in s 2 2) i to n o su rfa c es of fi n i te a 3 a n d f u r th e rm o re t he /n i te u c a i ns d o t V r re m r i f fi p infl a te } In other cas e s special tes ts m u st b e , be as x lwa ys th a t th e so certain whether the formul a can or the more c omplicate d one ( 35) mus t be re s o rte d to a a Fo r e t e ns i o ns a nd m difi o ca i i . t o ns o f t h s t h e o re m , se e e xe rc iws . us e d VEC TOR 222 A NA L YSIS irregul i ies which may arise are th own into the inter not into the analy tic appearance of the formul e his is the essence of Professor meth d of treatment Th e re p T ar t r t a ti o n , a G ib b s ’ s o . . Th e fi rst partial derivatives of the potential may also be obtained by differentiating under the sign of in te gration 1 . d x2 d y a d z z ( 36 ) w. V[< a like mann er for a vector function In Pot W 9 — 9 x W W x ( 2 1) 517 , y2 , z z ) 1 V P ot V= i 9 Pot V 9 Pot V 9 a 1 wi l l, k (z 2 3 s B ut m 1 r] doe 2 dy“dz “ If a n a n n e r, i ( a tt e :c 2 mldp it wo u — x1 ) t we re + md be se e n j( y g — y 1) e to o k (z mi g bta i n t ha t t h e vo l u a + the e g g pa rt i a l d e ri va t i ve s in t h e ra ls n o l o n e r c o n ve r e d se c o nd n te z . sa m e S IN TEGRAL CAL C UL U OF VE CTOR TH E Henc lik In e 22 3 III V Pot V e S mann e r W V x Pot V o P ot III W= fff il These th ree integrals obtained from the potent a differentiating O perators are of great importance in mathe m a t ic a l physics E ach has its own interpretation Co n s e quent y although obtained so simply from the potential each M oreover inasmuch as these is given a separate name integrals may ex ist even when the pote ntial is d vergent they must be considered independent of it The are to be looked upon as three new integrati ng operators defi ned each upon its own merits as the potential was de fi ned et therefore l . . i y . . L , . , , r 1 12 ' 12 X ° W ( 39 2 g a z , w( (13 2, y 2a z 3 If the p o te n ti a l e xis ts , e 2 ) d xg d y g d z >d x 2 dy2 dz z 2 L ap 43 ( ) W = Ma x w . then V P o t V = Ne w V W La p W P o t W = Ma x W VV X P ot Th e fi rst is writ en t . N e w V and read “ Th e N ewto ni an of ( 44 ) VE CTOR 224 A NA L YSIS reason for calling this integral the Newt n is t t repre ent the density a body the integral give the f due to the bod y at the point ) ( f wi be proved later second is written r ad Lap acian of W his integral was used to a c n iderable tent by L aplace is of fre quent occurrence in electricity and magnetism represent the current in space the L aplacian of gives the at the f oint ( due to the c rrent th i rd is w tten ) and read the Ma wellian of W his integral used by Ma we l too occurs freq ently in electricity magnetism instance if represent the inten ity o n ia Th e V a o o s x 1 s l m a gn e tic 0 p pl , z u l Ma x W l and . It , . orc e Th e . ri wa s T ” x x d If W . x , l an It . 0 Thi s . T . ex s , if o rc e La p W ” l th e pl , Th e . “ o s ttra c tio n ll e f ha . u , Fo r W s of magnetiz ation I the M axwellian of I gives the m a gne tic p o te n ti a l at the point ( 501 p l s l ) due to the magnetiz ati on To show tha t the N ewtonian gives the force of attr actio n acc ording to th e la w of t h e inverse s ua e of th e distan ce et d m z be an e ement of mass situa ted at the point x T h force at e ( 2 y, (x1 y1 z l ) due to d m s e ua t o , , L , . q r y l , , , . i q l , r in magnitude and h as t h e di e cti on of the vector rm from the po nt ( x1 9 1 2 1 ) to the point (mg 31 2 z z ) enc e th e fo ce is i , , , Integrating , . H . ll r over t h e entire body or over a spac e a cco to the conventi on here adopted the total force is , , r 12 V 31 , 4 7 r denot s the den ity of matter whe e V e s . d v2 = Ne w V r din g VEC TOR 226 Integrating the point A NA L YSI S over all space the total magnetic force acting z ) upon a unit positive pole is gl l , , tl v L ap W j L et ’ This integr a l may be ex panded in terms of i W ( x2 z , a i X Q’ s i ) z z) j1 + k x Th e i, j 2 . , k 1 ) L components of ap (ya 9 1 z , a ) Z ( “ z _ z 2 , z z z ' — z . 19 2 , 2, are respectivel W — 3 ( k , 0 7 Z ( ( ya i + , k ) 1 a ) z ) . y Y ) 1 r j La p W k t — x 1 La p W ) Y — (y 2 — g1 ) X 3 y erms the potential (if one ex ists ) thi s ma be i La p W ” La p k . La p W 9 Pot V 9 a yl a 9 1 9 Pot V 9 50 , 1 9 F ot z 9 Po t X 9 a x 1 9 Pot X y i be the intensit of magnet z ation at the point ( x 2 v2 z 2 ) that is if I be a vector whose magnitude is equal to the magnet ic moment per un t vo ume an whos e To show that if W 9 Fot z , , , I , i l d IN TEGRAL CAL CUL US TH E VECTOR S OF 2 27 direction is the direction of magnetiz ati on of the element d v2 fr om south pole to nort h pole then the Max wellian of I is th e magnetic potential due to the di stribut ion of magneti z ati on Th e magnetic moment of the element of volume d v2 is I d v2 Th e potential at (513 1 p l s l ) due to thi s element is equal to it s magnetic moment di vided b y the square of the distance r 1 2 and multi plied b y the cosine of the angle b etween the dire c tion of magnetiz ati on I and the vector rm Th e potential is therefore , . . , , . Inte g rati g the total magneti c potential is seen n , Ma x I d vz l Let 0 0 , 2 . integral may a so be writt n out in term Thi s 1 to b e V 2, z a 11 4 ) ( as , 02 0 — x e z , 1 ) ( V2 “ , k 0 002 B z) +j A of x s V1 ) ( B z — z 1 z ) . y, 9 2a z z . z ) C . inste ad of se , Z/v z 1 the variables x g z ; and inste ad of 1 xpression takes be used the e x z the variables C n y 5 2 2 2 Maxwell o n the form g ven b If , , i , Ma x I , y , , , . A (g ’‘ ing to the notation employed for the Laplacian Accord Ma x W 1 Ma xwe ll : El e c t ri c it y and Ma gn e t is m , Vo l II . . p . 9 . VECTOR 2 28 q y Maxwell ian of a vector fun ction is a scalar uanti t ma be wr tten in te rms of the pote ntial ( if it e xists ) a s Th e It A NA L YSIS y i Ma x _ W 9 Pot X 9 x 9 Pot Y . 9 Fo t Z + 9 1 z r Thi s 1 y form of expression is much used in o dinar t reatises upon m at hematical physics aplacian and M axwellian however should Th e N ewtonian not b e associated indissolub ly with the particular physical interpretations given to them above They should be looked upon as integrating operators which may b e applied as the potent ial is to any functions of posi t ion in sp a ce Th e N e w t o n ia n is applied to a scalar function and elds a vector function Th e aplacian is applied to a vector fun ction and ields a fun ction of the same sort Th e Max wellian is applied to a vector function and ields a scalar functi on Moreover these integrals should not be look ed upon as the derivatives of the potential If the potential e xists the are its derivatives B u t they frequently e xist when the potential fai ls to converge et V and W b e such f u nctions that their potentials 9 L] e xist and ha ve m general de fi nite values Then by ( 2 7) and , L . , , , . , yi , y L . y . . . y , . . L . . ( 29 ) V V Pot V : o B ut by ( 45) V Pot V V= o V Pot V and : Pot V N e w V, o V o V Pot V= V o Ne w V = Ma x V V Pot B y ( 2 7 ) and ( 2 9) V V ut VV V P o t V V = Ma x V V Hence B o b y ( 4 5) ' and by (45) V V Pot VVV W= V P o t V W Pot W Ma x W V W Ne w V W Pot : , z . 4 6 ( ) P ot V V W . VE CTOR 230 P o is s o n 92 ] Le t V be h a s in ge ne ra l a a ny fi de 9 P 0t V 9 33 E qu a tio n s fun c tio n i n sp a t he p o te n tia l c e su ch P ot V n ite va V 2 ’ A NAL YSIS -V lu e Th e n . Pot V 92 P ot V — 9 2 Pot V 9 2 1 Thi s 4 vr V, z 4 2 qr V 1 q equation is known as Poisson s E uation Th e integral which has been de fi ned as the p o te n tia l is solution of P oisson s Equation Th e proof is as follows ’ . a ’ . . Pot V V r V V 12 2 3 subscripts I and 2 have been attached t o designate clearly what are variables with respect to which the di fi e re n t ia t io n s are performed Th e ~ . Vl and v, o P o t Vl V v, o Ne w V Vs v z . TH E IN TEGRA L A LCUL US OF VECTOR C v, v1 or V2 V = V V 2 -V2 + v2 S 231 V v, VV . Integrate V2 V d v 2 That He e h is to say VI s i L sat s fi es aplace s ’ Vl P o t V E quation . An d y b ( 8) V2 Vd v2 ( 5 3) l surface integra is taken over the surfac e which boun ds the region of integration of the volume integral This is taken over all space Hence the surfac e integral must b e taken over a sphere of ra di us R a large quantity and R m us t be allowed to increase without limi t At the point (x1 y 1 z l ) however the integrand of the su rface integral becomes in fi n i te owin g to the presence of the term Th e . ” . , , , , 1 , , VEC TOR 23 2 A NA L YSIS Hence the surface S must include not only the surfa ce of the Sphere of radi us R but also the surface of a Sphere of ra dius ’ R a small quantity surrounding the point ( x1 p 1 z l ) and R must be allowed to approach z ero as its limit As it has been assumed that the potential of V e x ists it is assumed that the conditions given (Art 8 7) for the e x istence of the potential hold That is , ’ , , , , . , . . Vr K when a X, Vr when is large r , r is small Introduce ( x1 , y1 , . i polar co ordinates wi t h the orig n at the point c ) T hen r becomes simpl r l y . 1 r Then r 3 r 12 for the large sphere of radius R VI 1 r z s in O d O d cp . Hence the surface integral over that sphere approaches z ero as its limit Fo r . 2 7r da Ff l Hence when R becomes infi nite the surface integra over the large sphere approaches z ero as its limit Fo r the small Sphere . Vl 1 o d a r z s in ’ d db dd ) . Hence the integral over that sphere becomes V s in O d d d d ) . VECTOR A NA L YSIS 2 34 Theorem : If V a n d W a th a t the ir p o t e n ti a l s V W a nd P o is s o n ’ s fi a re re s u c h f unc ti on s fp o si ti o n i n sp a c e ll p o i n ts a t wh ic h o in ge n e ra l, th e n fo r a a n d c o n t in u o us th o s e p o te n t ia ls e xis t n i te sa tisfy E qu a tio n , V V Pot V V Pot W V 2 5 ( ) modifi cations in thi s theorem w h ich are to be made at points at which V and W become dis c o n tin u o u s will not be taken up here It was seen (4 6 ) Art 9 1 that Th e . . V V P ot V V V : o -Ne w V — Ma x V V In Ma x V V Ne s 4 4 vr vr V V . a similar man ner it was seen ( 51 ) Art 9 1 tha t VV P o t W = V Ma x W— V > < La p W — Ne w V W . ~ V Ma x Ne o By W — V x La p W W — La p V x W li vi rtue of this equa ty W Lap x V W where and W1 W2 is WI Lap x W V N ew V W — di vided W 4 — 7r 4 W 54 ( ) , 7r W . into two parts Ne w V W 5 5 ( ) L ap curl ( 56 ) . W2 . , N ew W div W ( 57) . THE INTEGRA L E quation AL C UL US OF VEC TOR C S 2 35 ( 55) states that any vector fun ction W multiplied L by 4 71 is equal to the difference of the aplacian of its c ur l and the N ewtonian of its divergence Furt hermore . -W1 V B ut the o V -W1 div W1 1 V > < W2 ut WI ivergence of the curl of a vector function is d Hence B VV x La p V La p V x W = 0 ero 4 ” V X V Ma s i the curl of the derivative of a scalar funct on is z ero Hence Conse V xW c u rl W2 2 quently any vect r fu ction . 5 8 ( ) 1 -W2 z . O . . ( 59) . whi ch has a potential may b e di vided into two parts of which one has no divergence and of which the other has no curl This division of W into t wo such parts is un ique In case a vector fun ction has no potential but b oth it s curl and divergence possess potentials the vector function may be divided int o th ree parts of which the fi rst has no divergence the second no curl ; the third neither di vergence n o r curl o W n . . , , L et W La p V x W V . , Ne w V -La p V x W V x Ne o W Vo — — 1 4W i — W l—W3 . W= 0 Pot V X V Pot V l W= O . d vergence of the fi rst part and the cur of the secon part of W are the refore z ero Th e . d VECTOR A NAL YSIS 23 6 x La p V V x W W P ot VV - VV V W P ot o V x Pot W V Pot W V x V . W= O V Pot V o -V x , W= O . Hence Hence 1 4 ' 71 l q cur of W is e ual to the curl of the Th e 1 4 7r fi r st part La PV XW into whi ch W is divided Hence as t h e second c u rl the third part can have none Moreover . , p rt has a . 1 4 W1 W = V -W= V - V Ne o 7r the divergence of W is equal to the the second part Thus di vergence of W N ew V i . . into which W is d vided Hence as the fi rst part has no divergence the third can have none Conse uently the thi rd pa rt W8 has neither curl nor divergence This proves the statement B y means of Art 9 6 it may be seen that an function W3 which po ssesses neither curl nor divergence m u st either q . . . . . y , VECTOR A NA L YSIS 238 With re sp e c t to scalar fu n c tio n a V o i dv or a re a nd Ma x a n d a lso in ve rs e Op e r a to rs V th e op e ra to rs 1 4 a nd 7 V . Ne w V : V Ma x V V = V . Wi th re sp e c t to a solenoidal fu n c tio n W1 1 4 a re Pot a nd 7r in ve rse op e ra to rs . That o th e r c url c url is Pot Wi th re sp ec t to a n a re in ve rs e op e ra to rs 1 — 4W Wi th O e ra p to rs . Pot V V res e c t p to W1 l irrotationa fu n c tio n W2 Pot That w, a n y Op e ra to rs a nd — th e = W1 ( 64) . op e ra to rs V V is — vv Pot w2 = W2 scalar or vector fu n c tio n V, W the NTEGRA L TH E I 1 4 Op e r a I7r P ot V re sp e c t t h e f to r s V V o to a VW V -V 239 Pot V= V v v — Pot o solenoidal fu n c tio n W1 W= W ( 66) . t he defe re n ti a ting s e c o n d o rde r V a re e qu i va — S " ‘ With AL CUL US OF VECTOR 71 1 — Pot V C V V x V X a nd le n t ( 67 ) Wi th t to r esp ec i n g Op e ra to rs an f th e irrotational fu n c tio n W2 se c o n d o rde r o V a re e u iva q By le n t V a nd is VV W2 V V That . i q i = V V 4 = V x La p 7r i 7r 4 Pot V 7r — 7r L lr 1 is a n op e ra fu n c tio n s rr op e ra to r W Pot — Ma x N e w V 9 6 ( ) W — P o t V Ma x W W — N e w M a x W ( 7 0) La p irrotationa l . ve c to r u n c tio ns f N ew Ma x to r wh ic h is e q u i va le n t to th e V Ma x La p a nd 4 — Ne w V W= P ot P o t W = ap sca a u n c t i o n s f f or al n Pot V P ot 4 enc e . V N ew V d W W by means of the potent al i tegr an H -WZ integra t n g the e uat ons 4 WV 4 the dife re n ti a t Pot L ap Lap . Fo r solenoi dl a vec to r VECTOR A NAL YSIS 2 40 i rs t Op e r a to r F or any ve c t o r fu n c ti o n th e f h e p o te n ti a l t i v es g a rt ; the se c o n d, the irrotational h e p o te n t i a l of th e i t v es p g o te n ti a l of th e solenoidal p a r t p . . There are a numb er of do u ble volume integrals whi ch are of such frequent occurrence in mathematical physics as to merit a passing mention although the theor of them will not b e developed to any considerable e x te nt These double integrals are all scalar quantities They are not scalar func tions of position in Space They have but a Single value Th e integrations in the e x pressions may be considered for convenience as e x tended over all Space Th e fu nctions b y vanishing identi cally outside of cert ain fi nite limi ts deter mine for all practical purposes the limits of integration in case t hey are fi n i te G iven two scalar fun ctions U V of positi on in Space Th e m u tu a l p o te n ti a l or p o te n ti a l p ro du c t as it ma be called of t h e t wo functions is the sex tuple integral y , . . . . . . y , , ” Pot ( U V) , On e y fff . , Wd vl d v2 . 71 of the integrations ma be performed ff P o t V d v, V ( :c 2 , y, , z Pot v U d ) , z 2 7 ( ) . l In a Similar manner t h e mutual potential or potentia product of two vector functions W’ W” is , z l) ° W I! d vl d e g . Th is l q y is a so a scalar uantit ried out . On e y integrati on ma be car VECTOR ANAL YSIS 24 2 yields O n e integration M a x Wat v l M a x (W, V) 8 7 ( ) A rt 9 3 3 5 ( ) By . 4 v . . UPot V= 7r P o t V] = ( V N w U e [ (V Ne w U) P o t V . V P ot V Ne w U) P o t V + ( Ne w U) - o . — = V t V P o N e w U) V ( P o t V ] + N ew U N e w V N w U e [ o o Inte grate U P ot Vdv + 47 ff — fff V Ne w U o d v P N t e w o V U ] [ Ne d v . fff Pot ( U V) = , Ne w U Ne d ff v P o t V Ne w U d _ a 9 7 ( ) . S surface integral is to be taken over the entire surface S bo undi ng the region of integration of the volume integral As t his region of integration is all Space t h e surface S may be looked upon as the surface of a large sphere of radius R If the functions U and V va n is h identicall y for all points out side of certain fi nite limits the surface integral must vanish Hence Th e . ” , . . , ( my ( 54) Art 9 3 . 4 n w ” o , Pot W ’ V x Lap W V Ma x ” W ” W Pot W ’ Pot ’ . TH E I B [ L ap W V ut and NTEGRAL CA L CUL US ” M a x [ V W] ’ X Pot W ” Pot W ap W W] W Ma x W ’ V ” H ence V x L ap W ’ V ” V Ma x and W Pot W ’ ” M a x W Ma x H ence substitut ng W] ’ ’ [ Ma x W V . Pot ” ” , ” [ L ap W x Lap W Lap W Pot W ’ ” W ’ W Pot W V Ma x ’ P ot ” V x Pot ” 24 3 x Lap W V ’ Pot L VECTOR S OF , W] ’ Pot W ” ’ i 4 7r Lap W L ap W [ L ap W x W Pot W ’ ” ’ Ma x ' ” V V M a x [ Pot (W W ) ff ff P ot W ” W] ’ Pot Integrating 4 7r ’ ” , Pot W x La p W ” ’ da — La p W ’o La p W M a x W Ma x ’ ff ” W dv ” dv ( 8 0) Ma x W P o t W d a ’ ” o . . W and W e x ist only in fi ni te s p ace these surface integrals taken over a large sphere of radius R must vanish a n d then If now 4 7: ’ Pot There ” fff W L ap W dv ff W Ma x W dv = La p Ma x ’ ” ’ ” . are a number of useful theorems of a function theo re tic nature whi ch may perhaps be mentioned here owing VECTOR A NAL YSIS 244 i l to their intimate conn ection w t h the integra calcul us of vecto rs Th e proofs of them will in some instances be given and in some not Th e theorems are often useful in prac t ical applications of vector analysis to physics as well as in purel mathematical work Th e o re m : If V ( x y z ) be a scalar function of position in space which possesses in general a de fi nite derivative V V and if in an portion of space fi n ite or infi nite but necessarily continuous that derivative vanishes then the fun ct on V s constant throughout that portion of Space . y . . , , y i , , , i . V V= O . V Choose a fi x ed point ( 901 : cons t . ) in the region , y ,, z , By . 2 ( ) p ge a 1 80 VV y f f dr = V V- Hence . If V (x , y, z z V a l , ?/p ) z 1 ) dr = 0 0- V ( x1 , y 1 , V ( v, y , z ) Th e o re m , s const l) . ) be a scalar function of position in sp a ce which possesses in general a de fi ni te derivative V V; if the divergence of that derivative e xists and is z ero through out any region of Space } fi nite or infi nite but necessaril continuous ; and if furthermore the derivati ve V V vanishes at every point of any fi ni te volume or of any fi nite port ion of surface in that region or bounding it then the derivative vanishes throughout all that region and the function V re duces to a constan t b y the preceding theorem y , m g . m Th e t e r thro ugho ut a ny re io n of s a c e us t b e g p bo un da rie s o f t h e re i o n a s we ll a s t h e re gi o n i ts e lf 1 . re g d d as ar e in c l u din g t he VE CTOR ANAL YSIS 24 6 an e x tension of the reasoning V is seen to be constant throughout the enti re region T The o re m If V ( x y z ) b e a scalar function of position in space possessing in general a derivative V V and if through 1 out a certain region T of Space fi nite or in fi nite continuous or di scontinuous the divergence V V V of that derivative e xists and is z ero and if furthermore the function V p o ss e s s e s a constant value 0 in a ll the surfaces b ounding the re gion and V ( x y z ) approaches 0 as a limit when the point ( cc y z ) recedes to infi nit then throughout the entire region T the func t ion V has the same constant value c and the derivative V V vanishes Th e proof does not di ffer essentially from the one given in the case of the last theorem Th e theorem ma be gen e ra liz e d as follows Th e o rem : If V ( x y z ) be any scalar function of position in space possessin g in general a derivative V V; if U ( x y z ) b e any other scalar function of position which is either posi tive or negative throughout and upo n the b oundaries of a region T fi nite or in fi nite continuous or discontinuous ; if t h e divergence V [ U V V of the product of U and V V ] ex ists and is z ero throughout and upon the b oundaries of T and at in fi nity ; and if fur t hermore V b e constant and equal to c upon a ll the b oundaries of T and a t infi ni t y ; then the function V is constant throughout the entire reg on T and is equal to c Th e o re m : If V ( x y z ) be any scalar function of position in Space possessing in general a derivative V V; if through out any region T o f space fi nite or infi nite continuous or dis con t inuous the di vergence V V V of this derivati ve e x ists and is z ero ; and if in a l l the b ounding surfaces of the region T t h e normal component of the derivative V V va n is h e s and at in fin ite distances in T ( if such there be ) the p roduct . , , , , o , , , y , , , , . y . , , , , , i . , , , , , 1 Th e re gi on i n c l u de s it s bo un da ri e s . , NTEGRAL CAL C UL US TH E I OF VECTOR S 247 vanishes where r denotes the distance measured from any fi x ed o ri gin ; then throughout the entire region T the derivative V V va n is h e s and in each continuous portion of T V is constant although for difi e re n t continuous portio n s this constant may not b e the same This theorem may b e generali z ed as the preceding one wa s b the substitution of the relation V U V V) O for r 2 9 V/ 9 r , , . y VV V= Oa nd Ur 2 9 V/9 r = 0 fo r r 2 9 V/9 r = 0 . corollaries of the foregoing th eorems t h e following statements ma be made Th e language is not so precise as in the theorems themselves but will pe rh aps be under stood when they are borne in mind If V U = V V then U and V di fi e r at most b a constant f V U = V V in any fi n ite If V V U = V V V and port ion of surface S then V U V V at all points and U difi e rs from V only b y a constant at most V V and if U V in a ll the bounding If V V U= V surfaces of the region and at in fi n ity (if the region e x tend thereto ) then at all points U and V a rc equal V V V and if in all the bounding surfaces If V V U of the region the normal components of V U and V V are 2 9 equal and if at infi nite d stances r ( U 9 r — 9 V/9 r) is z ero then V U and V V are equal at all points of the region and U difi e rs from V only b y a constant ’ ” Th e o re m If W and W are two vector functions of position in Space which in general possess curls and divergences ; if for any region T fi nite or in fi n i te but necessarily conti nuous ’ the curl of W is e ual to the curl of W and the di vergence ’ of W is equal to the divergence of W ; and f moreover ” the two functions W and W are equal to each other at every point of any fi nite volume in T or of any fi n i te surface ’ ” in T or bounding it ; then W s e ual to W at ever point o f the reg on T As y . , y . , i . o o , . : . , i , . , q , ” ” i ’ i i q . y 24 8 xW VECTOR A NAL YSIS A vec W V x (W 1 h a t e tor function whose curl vanishes is equ l to derivative — = V V W W of a scalar function V ( page et Then V V V O owing to the equalit of the divergences Th e theorem therefore b ecomes a corollar of a precedi ng one Th eo re m IfW and W are two vector functions of pos i ti on which in general possess de fin ite curls and divergences 2 if throughout any p e rip h ra c tic region T fi nite b u t not of n ecessarily continuous the curl of W is equal to the c u r W and the divergence of W is equal to the divergence of W ; and if furthermore in a ll t h e b oundin g surfaces of the region T the t a nge n ti l components W and W are e ual then W is equal to W throughout the a p e rip h ra c t ic region T Th e o re m : If W and W are two vector functions of posi tion in space which in general possess de fi nite cur s and divergences ; f throughout any a c yc lic region T fi nite bu t not necessarily continuous the curl of W is equal to the curl ’ W and the di vergence of W is e ual to the divergence of W ; and if in a ll the b ounding surfaces of the region T the d l components of W and W are equal ; then the fu nc tions W and W are equal throughout the region acyclic T Th e proofs of these two theorems are carried out by me a ns of the device suggested before The o re m : If W and W are two vector functions such that V V W and V V W h ave in general de fi nite values in a certain region T fi nite or in fi ni te contin uous or discon ti n u o us ; and if in ll the bo unding s u rfaces of the region and at infi nity the functions W and W are e u al ; then W is equal to W throughout the en t r e region T Th e proof is given by t e a tin g separatel t h e th e e c o m p o n e n t s of W and W Since V ’ = V x ’ ” , L y y ’ ” . . . ” ’ a l , ’ , ’ ” ” ’ a ’ q ” ” . ’ ” l i , ’ q , ” ” ’ ” ” ’ . . ’ ” ’ ” , , q a ’ ” r ’ ” i y ’ r . ” . 1 Th e A ti c . re If it ggi i re on on l Tm wh i c h ay h a ve t o b e gi e n c l o se s e n c o se s n o re on wi t h it is mi d f y li by gi i a e ac n i ts e l a p e rip hra c c c the a n o t h e r re t ic . i n se rt o n o f on is sa id t o dia ph ra gm s . b e p e rip h ra c VECTOR A NAL YSIS 250 fff — ff 8 d c + Z d wde l. d e + d d X y [ if X Y Z be the three components of the vector f u nction W Stokes s Theorem : Th e surface integral of the curl of a vec t or func t ion taken over any surface is equ al to the line integral of the function taken around the line boun d ng the surface An d convers ely if the surface inte gra l of a vector func t ion U taken over any surface is equal to the lin e integral of a function W taken around the boundar then U is the curl of W . , , ’ i . y , . W V X a nd if ff U -d a S s ‘ da W t dl ' , ( 1 2) 0 Application of the theorem of Stokes to deducing the equations of the electro -magnetic fi eld from two e xperimental facts due to Faraday Application of the theorems of Stokes and G auss to the proof that the divergence of the curl of a vector func t io n is z ero and the curl of the derivative of a scalar function is z ero Formul a analogous to integration b parts . y . q ff V S ff s u X V Vu v v v o da u v -d r — ff u -d r q , v S da q -d r — ; j vV u d r ( 1 6) , VECTOR A NA L YSIS 25 2 V -V P o t P ot V -P o t W = P o t V V -V W 31 ( ) -W 3 2 ( ) VV , , 3 3 ( ) Th e integrating operator Pot and the differentiating operator V are commutative Th e thr ee additional integrating operators known as the N e wto n i a n the La p la c i a n and the M a xwe lli a n . . , , Ne w V Ma x W If the i rl i potent al e xist s these integ a s are related to t as fol lows : V V Pot V: N e w V, W = La p W; P o t W = Ma x W x Pot V . interpretation of the physical mean ing of the N ewtonian on the assumption that V is the density of an attra cting b ody of the aplacian on the assumption that W is electric flu x of the Max welli an on the assumption that W s the intens it of magnetiz at on Th e e x pression of these integra s or their components in terms of ac y z formul a and Th e L , , y i i . , V o , N e w V = M a x V V, V Ma x W = Ne w V -W , l VE C TOR A NAL YSIS 2 54 1 4W Ne w V = V Ma x V V = V . 4 6 ( ) P ot V VWZ P ot V P ot V — -V V V — VW — V Pot 1 -V V V 4 W2 = W2 ( 65) . P ot V 7r 1 -V -— - P o t W = W ;r 1 £ . ( 6 7) VV 4 4 7r 7: W2 Pot V Pot W M u tu a l pote ntials N -W2 ( 6 8) M a x Ne w V ( 6 9) = VV L ap L ap W W ( ) ewtonians L aplacians and Ma well an Ne w M a x x , , 70 . i a may b e formed They are sex tuple integrals Th e integra tions cann ot all b e performed immedi ately ; b u t t h e fi rst three may be Formul a ( 7 1 ) to ( 8 0) inclusive deal wi th these inte grals Th e chapter closes with the enunciation of a number of t heorems of a function -theoretic nature B y mean s of these theorems certain facts concernin g functi ons may be inferred from the conditions that satis f aplace s equati on and have certain b oundary conditions A mong the e x ercises numb er 6 is worthy of especial atten tion Th e wo rk do n e in the te xt h a s fo r th e m o s t p a rt a s s u m e d . . . . . yL ’ . . t h a t th e p o te n tia l N e wto n ia ti a l do e s n s, e xis ts La p l a c i a n o t e xis t . . B ut n s, a nd Th ese ma f ny o M a xwe lli a th e ns o r m u l ae c o n n e c tin g f h o ld whe n th e p o te n are taken up in Ex ercise 6 referred to . NTE GRA L CAL CUL US TH E I EX ER CISE S OF VE C TOR S 55 CHA PTER IV ON is a Scalar function of position in space the line If V integral q is a vector uantity Vd Show . that Vd r d a a > <VV . That is the line integral of a scalar fun ction around a closed curve is equal to the s ke w surface integral of the deriv ative of t h e function taken over any surface spanned into the contour of the curve Show further tha t if V is constant the in t egra l around any closed curve is z ero and conversely if the integral around any closed curve is z ero the function V is constant Hint : Instead of treating the integral as it sta nds m u ltiply it ( with a dot) by an arb itrary constant un it vector and thus reduce it to the line integral of a vector function 2 If W s a vector function the lin e integral . . i . . d q r is a vector uantity It may be called the s ke w line integra l of the function W If e is any constant vector Sho w that if the inte gral b e taken aro u nd a closed curve . , . ( V c Th e 1 we ll ’ s c he fi W— fo u r e xe rc is e s a re Th e o ri e de r El e c t ric ita t rs t ta c -V W) -d a k f ym en ro wh e re t h e P o p pl a re ’ d kd wo r r, Ein fu h run g in die M a x s e o ut . VECTOR A N ALYSIS 2 56 and H o = c c V o Wd a W [c e (W V o d a ) o case the integral is taken over a plane curve and the surface S is t h e portion of p ane included b y the curve l In [ VWd a — V (W i Show that the integral taken over a plane curve van shes when W is constant and conversely if the integral over an plane curve vanishes W must be constant 3 Th e s u rface integral of a scalar function V is y . . Vd a is is a vector quantity . that the surface integral of V taken over any closed surface is e ual to the volume integral of V V taken throughout the volume boun ed b that surface That is Th . Show q d y . V Vdv . r Hence conclude that the surface integral over a closed su fac e vanishes if V b e constant and conversely if the surfac e integral over any closed surface vanishes the function V must be constant 4 IfW be a vector function the surface integral . . , da > <w may be called the s ke w surface integral It is a vector quantity Show tha t the Skew surface integral of a vector . . VECTOR A NAL YSIS 2 58 By e xercis e ( 3) V2 ( p 1 2 V) d v2 p Vda 12 . can be Shown that if V is such a function that N e w V e xis t s then this surface integral taken over a large Sphere of ra di us R and a small sphere of radius R approaches z ero when R b ecomes inde fi n itel great ; and R inde fi nitel small Hence It , ’ y y ’ , . P Pot Ne w V : 12 V2 V d vz r VV Prove in a Sim ilar manner that W= P o t Ma x By means of -N ew V Ma x fff = p V Ma s enc e Hence 7 . An W . r ( 8 7) it is possible to p ove V H V La p W— V > < La p 12 V = Ma x V V, W= VV o Ne w V W d vz fff p V Ma x W— integral us ed by — . fff pl Vd l W — V Ma x H elmholt r W fff p W= 4 z 12 z 7r is V d vz , 12 W . o Vd z . V V W d vz ° ’ S THE INTE GRAL CAL CUL U OF VECTOR S 2 59 or if W be a vector function 7 Show l that the integra converges if Vr 5 y VH H ( W) V XH ( W) =H ( o V) = H ( V VV H o (V V ( . W) =E ( = La p Z T ( Z 7 , W) 2 , W) , V V) = 2 P o t V (V Pot Pot H V) Ma x ( r W) V V) = M a x ( W) so rapi dl y V dim in is h e s = = N H e w V V V ) ( ( ) = H V V VE ' 9 0 ( ) . K when r becomes inde fi n i tel great o 12 W d vz V Pot Pot W ( W) 9 ( 7) . 9 8 ( ) a proof of G auss s Theorem which does not depe nd upon the physical interpreta tion of a func t ion as the flu x of a fluid Th e reasoning is similar to that emplo ed in Art 51 and in the fi rst proof of Stokes s Theorem Show tha t t h e division of W into two parts page 2 3 5 9 s unique 1 0 Treat in a manner analogous to that upo n page 2 2 0 the case in which V has curves of s continuit es 8 G ive . ’ y . . ’ . i , . , . . , di i , . H C AP TER V LINE AR VE CTOR FU N CTION S AFTER the de fi ni tions of products h a d be en laid down and applie d two paths of advance were open O n e was differential and integral calcul us ; the other higher algebra in t h e sens e of the theory of linear homogeneous substitutions Th e treatment of the fi rst of these topics led to new ideas and new symb ols to the derivative dive rgence c u rl scalar and vector potential that is to V V V x and P o t w th the au xiliaries the Newt onian t h e aplacian and the Maxwell ian Th e treatment of the second topic will likewise introduce novel t y b oth in concept and in notation the linear vector function the dyad and the d adic with their appropriate symboliz ation Th e simplest e x ample of a linear vector function is the ’ product of a scal ar constant and a vector Th e vector r . , , . , , , L , , , , , , i , . , y , . . 1 ( ) is a linear function of more g neral linear funct on may be obtained by consideri g the components of r indi vi d Let jk be a system of a es components of A r . i e n ua ll y . i, x , r a re i -r j o , k r, Th e . -r . L et each of these be mul iplied by a scalar constant which t may be different for the di fferent components 0 1 i 0 r, C z j ' r’ 08 k . ro . 2 62 fi VE C TOR A NA L YSIS vector r is said to be a linear vector fun c tion of another vector r when the components of 1 along three non-c oplanar vectors are e xpressible lin early with scalar coeffi cients in terms of the components of r along those same vecto rs De ' A n i tio n : . " . yb se a a ’ z z c w l where [ a b c ] , bl y ba y e x ba y c a g z , l e “2 x 0, z , z a z , then r is a linear function of r ( Th e constants a l bl c l etc have no connec t ion with the components of a b 0 par allel to i k ) A nother de fi nition however is found to b e more convenient and from it the foregoing may be deduced De n i ti o n : A continuous vector func t ion of a vector is said to b e a lin e a r vector function when the function of the sum of any two vectors is the sum of the functions of thos e vec t ors That is the function f is linear if ’ . . , , j , , , , , , . fi . , . f( r r 1 2 ) fOi ) 4 ( ) ' : be any positive or negative scalar and if f be a linear function then the function of a times r is a times the fun c t ion of r Th e o re m If a , . f Ca r) An d a hence f< 1 r l a a a z r m) “ z r f ( 2)+ z a r a e 3 . ‘ l ‘ f( r3 ) Th e proof of this theorem which appears more or less obvious is a t rifle long It depends upon making repeate d use of rela t ion . f< r ) r =f ( ) r + f ( r) me . L INEA R VE C TOR FUNC TI ONS Hence In like manner f ( 2 r) 2 f ( r) f ( n r) n . f ( r) where n is any positive integer et m be any other positive integer jus t obtained L . f( ) r Hence =f 2 63 Then . y b relation m n n r f( ) . That is , equation ( 5) has b een proved in case the constant is a rational positive num b er To show the relation for negative numbe s note that r . = 0 f( ) f( 0 0) = 2 f ‘ i d» — = 0 r f( ) f < ) r o . — = r > > f ( ) + f( = f< r f <> — r a — r — fc e ) r . prove ( 5 ) for incommensurab le values of the cons tant it becomes necessary to make use of the continuity of the a function f That is To , . f( L et = ) f ( a? r ) 90 r approach the incommensurable n umb e r a through a suite of commensurab le values Then a: . f( H ence ) x r r w f( ) a f( ) r passing VECTOR A NAL YSIS 264 ( a: r r f( ) Hence ) a a f ( r) which proves the theorem Th e ore m : A linear vector function f ( r ) s e ntire mined when its value for three non -coplan ar vec t o s a known i . Let Since ly det r r are e , . b, c . l = f ( a ), r m f I =f . b ( )’ is any vector whats oever it ma , = r x a + yb + y presse be ex . e xp re s s e d i a particular case of a l inear funct on as i o l r c r i jz j fk os k i as z a Hence In Art 9 7 d o wa s r . r the sake of brevity and to save repeati ng the vector whi ch occurs in each of these terms in the same wa th s may b e written in the symbo l ic form Fo r In b y i like manner if a 2 be any given vectors and b v b 2 another set equal in numb er the e xpre s sion , , , , r ' a l b l o r + a 2 bé o r + a 3 bg o r + u ( 6) o is a linear vector function of r ; for owing to the distrib ut ive c haracter of the scalar product this function of satis fi es Fo r the sake of brevity r may be written s m relation b o lic a lly in the for m r ’ r ' y VE C TOR A NAL YSIS 266 multipli ed into r by direct or scalar multiplication Th e order of t h e factors O and r is important Th e direct product of r into O is . . r O= r ( o a b1 + 1 _r a b1 + 1 a r y the vectors r 2 b2 b aa -a 2 b 2 + r o a a b 3 9 ( ) + 3 are in general difle re n t M De n i ti o n : When the dyadi c O is multiplied in to r a s O r When r is multiplied in O as O is said to b e a p refa c t o r t o r r O O is said to be a p o s tfa c to r to r Evidentl fi Q; Q Q . , . , . A dy a dic O u s e d e i th e r a s a p refa c to r o r a s ve c to r r de t e rm in e s a lin e a r ve c to r fu n c ti o n of r a p t ” a / c to r os to a two li near vector functions thus ob t ained are in general different from one another They are called c o nju ga te linear vector func tions Th e t wo dyadics . Th e . . ¢ = a b1 + 1 a z bz ' l a ' g bg each of which may b e ob tained from the other by inter changing the an t ecedents and consequents are called c o nju Th e fac t tha t one dyadi c is the conjugate of ga te dy a dic s another is denoted by affix ing a sub script C to either , . ’ . r dyadic used as a postfactor gives the sam e esult as its conjugate used as a prefactor That is Th e o re m : A . fi De be e n i tio n : wh e n O r s o O r T r O r T r z s o $0 o An y qu a l when or when or t T y 0 ( 9) 12 two d a di cs O and for all values of for all values of o r fo r a ll T are said to r r , , va l u e s o f s a n d r . L INEA R VECTOR FUNCTI ONS 2 67 third relation is equivalent to the fi rst Fo r if the vectors O -r and T r are equal the scalar products of any vector 3 in t o t hem must b e equal A n d conve rs ely if the scalar product of any and e ve ry vector 8 into the vectors O r and T r are equal then those vectors must be equal In like manner it may be shown that the thi rd relation is e q u iva lent to the second Hence all three are equivalent Th e o re m : A d adic O is completely determined when t h e values Th e , . , . o , . y . . O O a, y b, $4 , where a b c are an three non-c oplanar vectors are known This follows immediately fr om the f a ct that a dyadic de fi n es a linear vector function If r a: a zc yb Or = Oc , , , . . , , y consequently two d adi cs O and T are equ a l provided equa tions ( 1 0) hold for th re e non -coplanar vectors and th re e non-coplanar vectors 8 Th e o re m A n y l inear vector function f may b e represented by a dya dic O to be used as a prefactor and b y a dyadic T w hi ch is the conjugate of O to b e used as a postfactor Th e linear vector function is completely determin ed when its values for thr ee non-coplanar vectors ( sa i k) are known ( page et r . , . , y j L = b. f O) ' f ( i) Then , a, k f( ) c , . the l inear function f is equivalent to the dyadic O a i b j c k, to be used as a postfactor and to the dyadi c T: to be us ed as a prefactor fo ) r . r r T O 2 68 VECTOR A NAL YSIS y i stud of linear vector functions therefore s identical with the study of dyadics De n i tio n A dyad a b is said to be mu ti plied b y a scalar a when the antecedent or the consequent is multipli ed b y that scalar or when a is distrib uted in any manner between the antecedent and the consequent If a a a Th e fi l . , ’ " . ( a a ( b) a a )b a (a b) (a ' a ) (a ” b) . dyadic O is said to be multiplied b y the scalar each of its dyads is multiplied b that scalar Th e is written a O or O a A y . . dyadic a O appl ied to a vector r either as a prefactor or as a postfactor yields a vec t or equal to a times the vector obtained b appl ing O to r th at is Th e y y ( Th e o re m utive . : Th e a o r a ( Or) . y i combination of vecto rs in a d ad s dis t rib Th a t is (a and This O) + b) c a c + b y c d q y follows immediatel from the e fi nition of e ualit of dyadi cs Fo r ( a c + b c ) o r and a b o r + a c o r Hence it follows that a dyad whi ch consists of two factors each of which is the sum of a numb er of vecto s may be multiplied out accordi n g to the la w of ordi nar algeb ra -e x cept t hat th e o rde r o f th e fa c to rs i n the dy a ds m u s t be r m a i n t a in e d . , y , 3 form the scalar coeffi cients of the correspondin g e ual If the coeffi cients be equal then obviously O r = T r q , da ds be . , for any value of r and the dyadi cs b y ( 1 0) must b e aq u a l Conversely if t h e dyadics O and T are e ual t hen b ( 0) q , s -O o L s r T o y , r o for all va lues of s and r et s and r each take on the Then i jk = O j i TO ii k i j . , , v lu e s . z o O i j k , z T i, j z k -O i o o o o , o O k j . o z T k j T k Ti, ' these quantities are precisely the nine coe ffi cients in the e xpansion of the dyadi cs O and T Hence the correspon din g 1 coeffi cients are equal and the theorem is proved Th u anal tic s t at ement of the equality of t wo dyadics can some t imes b e used to greater advantage than the more fundamental de fi nition ( 1 0) b ased upon the conception of the dyadic as de fi ning a linear vector function The o re m A dyadi c O may be e x pressed as the sum of nine dya ds o f which the antecedents are an three given non coplanar vectors a b c and the conse uents any three given non -00planar vectors 1 m 11 E very antecedent may be e xpressed in te rms of a b c ; and every consequent in terms of 1 m n Th e dyadic ma then be reduced to the form B ut . y . . , q , , , . , , O= a a a 1 As a c o ro de pe n de n t . No n e o ft h e 11 21 , 1 + a bl + a 31 + a a m m xp lla ry o f t h e y 31 C t h e o re ma y be a b m + a 0m a 82 + ev , 12) a re . y n a la 23 38 b n o n . m id e n t t h a t t h e n i n e dya ds re ss e d li n e a rly in t e r s o f t h e it is e m + e lz gz , , o t h e rs . ih L INEAR VECTOR FUNCTI ONS 2 71 This e xpressi on of O is more general than that given in It reduces to t ha t e x pression when each set of vectors a b c and l m n coincides with i k Th e o re m : An y dyadic O may b e reduced to the sum of th re e dyads of which either t h e an t ecedents or the consequents but not b oth may b e arb i t rarily chosen provided they be non coplanar et it be required to e xpress O as the sum of three dyads of which a b c are the consequents et 1 m n be any o t her three non -coplanar vectors O ma then be expressed as in Hence , , , , , j . , , , L . , L y . , . c ( a , 31 1 + , a az m + “ 32 0 1 , ( 1 6) like manner if i t be required to e xpress O as the sum of three dyads of whi ch the three non-COp l a n a r vectors 1 m n are the conse uent s In q , O where L Mm Ll a M= a N = a ll 12 ls a a a + a a + a N n, b a c , 81 gg b + a 0 , 32 23 b + a m 21 , 33 e x pression s for O are u n iqu e Two equal dyadics which have the same three non -coplanar ante cedent s a b 0 have the same consequents A B C these however need not b e non-coplanar A n d two equal dyadics which have the same three non -00planar conse uents 1 m n have t h e same three antecedents 1 02 ] De n i tio n : Th e symb olic product fo rmed by the jux ta position o i two vectors a b without the intervention of a dot or a cross is calle the in de te rmin a te p ro du c t of the two vecto s 9 and b Th e . , , , , , . fi . q , , , . d . , , r 27 2 VECTOR A NA L YSIS reason for the term indeterminate is this two products a h and a x b have defi nite meanings O n e is a cer t ain scalar the other a certain vector O n the other hand the product a b is neither vector nor scalar — it is purely sym b olic and acquires a determinate physical meanin g onl Th e product a b does not ob ey when used as an operator the commutative law It does however obey the distrib utive law ( 1 1 ) and the associative law as far as scalar mul tiplication is concerned (Art T he o re m Th e indeterminate product a b of two vectors is the most general product in whi ch scalar multiplication is associat ive Th e most general product conceivable ought to have the property that when the product is known the two factors a re also known Certa inl y no product could be more general Inasmuch as scalar multiplicati on is to be associative that s Th e o . Th e . . , y . . . i . , a ( a b) ( a a ) b ( a a b) ( d ’ a . ) it will be impossible to completely determine the vectors a and b when their product a b is given An y scalar factor may be transferred from one vector to the other Apart from thi s possib le transference of a scalar factor the vectors com posing the product are kn own when the product is known In other words Th e o re m : If the two indetermi nate products a b and a b are equal the vectors a and a b and b must be collinear and the product of the lengths of a and b ( taking into account the positive or negative sign accordi ng as a and b have re s p e c t ive l y equal or opposite directions to a and b s e ua to t h e ) product of the lengths of a and b . . , . ’ ’ , ’ , ’ ’ ’ . ’ i q l ’ VECTOR A NAL YSIS 27 4 indetermin ate product a b imposes five conditions upo n the vectors a and b Th e directions of a and b are fi x ed and lik ewise the product of their lengths Th e scalar product a b b eing a scalar quan t i t y imposes only one condition upon b b eing a vector quantity Th e vector product a a and b imposes three conditions Th e normal to the plane of a and b is fi x ed and also the area of the parallelogram of which they are the side Th e nine indeterminate products ( 1 2 ) of i 1: into themselves are independent Th e nine scalar products are not independent O nly two of them are different Th e . . x , , . , , . , . j , . . . i -i z k jj o z o k z l, and Th e n ine vector products are not in dependent either ; for and jx k i xj k xi two products a b and a x b ob tained respectively from the inde t erm inate product b y insert ing a dot and a cross b e tween the factors are functions of the indeterminate product That is t o say when a b is given a b and a x b are determined Fo r t hese products depend solely upon the directions of a and b and upon the produ ct of the length of a and b all of which are known when a b is known That is Th e . , , . , . if a b z a ’ b ’ , a o h 1 ( 7) x does not hold conversely that if a b and a b are known a b is fi xed ; for taken toge t her a b and a x b impose upon t h e vectors only four condi t ions whereas a b impos es fi ve Hence a b appears not only as t h e mo s t general product but as the mos t fundamental product Th e others are merely functions of it Their functional nature is brought out clearl b y the notation of the dot and the cross It , . y . . . fi L INEAR VE CTOR F UNCTI ONS 2 75 scalar known as t he s c a la r of O may be o b t a in e d b y inser t ing a dot b et ween the antecedent and c o n s e quent of each dyad in a dya di c Thi s scalar will be denoted I b y a sub script S attached to O De A n itio n : . . If ( 1 8) like manner a vector known as th e ve c to r of O may be o b tain ed b y inserting a cross bet ween t h e antecedent and con sequent of eac h dyad in O This vector will be denoted b y attac hing a sub script cross to O In . . ¢x If O 3 1 X b + 1 3 2 X b2 + 3 X b3 + 0 8 be e xpan ded in nonion form in term s of i , j k, , 2 0 ( ) “ ) 23 i + ( 4 a a ) j ( 13 -O j+ k o Ox = (j o O k + (i q O k o j ) i + k ( o a o — -O O i) j j ( 21 ) Ok, O i o ) k k — i c O k) j . e uations ( 2 0) and ( 2 1 ) the scalar and vector of O are e x pressed in terms of the coe ffi cients of O wh en ex panded in the nonion form Hence if O and T are two equal dyadi cs the scalar of O is equal to the scalar of T and the vector of O is equal to the vector of T In . , . O If T, From OE : and T5 Ox TX ( 2 2) . thi s it appears that O and O are functions of O uniquely determined when O is given They may s ometimes be ob t ained more conveniently from ( 2 0) and (2 1 ) than from and sometimes n o t ( 1 8 ) and , )( . 1 a nd A su fre e m g m m . b s c ript do t i h t b e us e d fo r t h e s c a l a r o fQ if it we re fro li a bilit y t o i s i n t e rp re ta t io n . i s uf fi c e nt ly dis t i n c t VE C TOR A NAL YSIS 276 a dic s D f y P ro du c ts o giving the de fi ni t ions and proving the theore ms concerning products of dyadics the dyad is made the unde r lying p rinciple What is true for the dyad is true for the dyadic in general owing to _the fact that dya s a n d dyadics ob ey the distrib utive law of mul t iplication Th e dire c t p ro duc t of the dy a d a b into the De n i ti o n dy a d c d is written ( a b ) ( c d) and is by de fi nition equal to the dyad ( b c ) a ll In , d . fi . : ( That a b) a t c l 2 ( 3) is the antecedent of the fi rst a n d the consequent of the second dyad are taken for the antecedent and consequent respectively of the product and the whole is multiplied b y the scalar product of the consequent of the fi rst and the antecedent of the second Thus the two vectors which stand together in the product , . ( a b ) (c d) are multiplied as they stand Th e other two are left to form a new dyad Th e direct product of two dya dic s may be de fi ned as the formal e x pansion ( according to the dis trib utive law) of the product into a sum of produc t s of dyads Thus 1f . . . u T= ( 0 1 d1 + c 2 d2 + c a d3 + o u ) ) O (o l 1 _ a + a + a Th e 1 b1 2 b2 a b3 d1 c l d1 + a °1 d1 + a -c l d1 + a ° p a re n t h e s e s ma y be b 1 l b 2 2 3 o o ' c c z d2 + 2 d2 a 1 b 1 -c s d3 + u b3 mi tte d in e a c h o ft h e s e t h re e e xp i re s s o n s . VECTOR ANAL YSIS 278 an O d T ( + o T+ O T ’ l Hence in genera the product W4 9 4 ( + a . rr f r q + m + q + i may be e xpa nded formally according to the d stributive la w Th e o re m Th e product of th ree dyad cs O T Q is associa tive Th a t is i . , , . q ly y and conse uent parentheses as ( 26) i either product ma be wr tten (D , ‘ g p g . without , proof consis t s in the demonstration of the theorem for three dya ds a b e d e f taken respective from the th e e d adi cs O T Q = = a b d c e f d d i b a c a e f e b o ) ( ( ) ) ( ) ( a h c d a f d b e e b t d e c) a f e ( ) ) ( ( ( ) Th e y , , r . : o Th e ly , , o : y o o z o o o . r proof ma als o be given by conside ing as o p e ra t o rs r { < T> O, T, -r >Q < o Q o r r ’ T) r = (OT {( O o r ’ r ” , T) -Q § . -r = ( T o Q) r = T OT [ d ( { . values of r . Conse (O . quently r] T o r O o ’ , = r ” r . . an d Q L INEA R VECTOR FUN C TI ONS 27 9 y theorem ma be e x tended b y mathematical induction to the case of any numb er of dyadics Th e direct product of any num b er of dyadics is associative P arentheses may b e inserted or omitted at pleasure without altering the resul t It wa s shown above ( 2 4 ) t hat Th e . . . Hence the product of two dyadi cs and a vector is associative Th e theorem is true in c a se the vector precedes the dyadics and als o when the number of dyadics is greater than two B u t the theorem is u n tru e when the vector occurs b e tween the dyadics Th e produc t of a dya d c a vector and another dya di c is not associative . . i . , , . L et a b y be a d ad of ( Henc e O r) ( O, T¢ O b r) a h = r d c ( ) ab ( b o ( r T) . and e d a dyad of a a o r) -c d a b T . d r c ( ) . results of this article may be summed up as follows Th e o re m : Th e direct product of any number of dyadics or of any number of dya di cs with a vector factor at ei t her end or at both ends obeys the distributive and associative a ws of multiplication parentheses may be inserted or omitted at pleasure B u t the direct product of any numb er of dyadics with a vector factor at some other position than at either end is not associative — parentheses are necessary to give the expression a de fi nite meaning ater it will b e seen that by making use of the conjugate dyadi cs a vector factor whi ch occurs between other dyadics may be placed at the end and hence the product ma b e ma de to as sume a form in which it s associat ve Th e l . L . i i y . 2 80 fi De VECTOR ANA L YSIS Th e n i tio n : a vecto r r and of a vector r respectively b the equations y . ( r d b) a x yd i products of a d a a b nto into a dyad a b are defi n ed s ke w a r b ( x r) , (r x a ) b x (a b) . skew pro uct of a dyad and a vector at either end dyad Th e ob vious e x tension to dyadics is Th e is a . x r i direct product of any number of dyad cs multiplied at either end or at b oth ends by a vector whether the multiplication be pe rformed with a cros s or a dot is associative B u t in case the vector occurs at any other position than the end the product is not associative That is Th wre m : Th e . . (e ( -T ) , OT) x (e ) o r = O s r x ( O s) =e r r o -O e O) ¢ Furthermore s o r x O an . d can have no other meaning than r o s x s, x s, T the e xpressions s, , VECTOR ANA L YSIS 282 Degre e s a dic s D f y u l li ty N f o o l y was shown ( Art 1 01 ) that a dyadic cou d alwa s be reduced to a sum of three terms at most and t his reduction can b e accomplis hed in onl y o n e way when t h e antecedents or t h e consequents are specifi ed In particular cases it ma be possib le to re duce the dyadic further to a sum of two terms or to a single term or to z ero Thus let It . , y . . O= l + bm + a c n y are coplanar one of the three ma be e xpressed in te ms of the other t wo as If 1, m, r n l = x m + yn . y i d ad c has been reduced to two terms If 1 m n were all collinear the dyadi c would reduce to a single term and f they all vanished the d adi c would vani sh Th e o re m : If a d ad c O be e x pressed as the sum of thr ee terms Th e , . y i y i , . O= l + bm + a o n of which the antecedent s a b c are known to be non-coplanar then the dyadic may b e reduced to the sum of two dyads when and only when the consequents are coplanar Th e proof of the fi rst part of the theorem has just been given To prove t h e second part suppose that the d adic could be educed to a sum of two terms , , , . y r . O = dp + e q and that the consequents l m n of O were non-coplanar ' Th s supposition leads to a contrad ction m n Fo r et be the s ste m reciprocal to l m 11 That is i , y , mx n l m n [ ] i , , . l . , . ’ , LINEA R VE C TOR FUNC TI ONS r 2 83 vec t o s l m n e xist and are non -coplanar be cause have been a s sumed to be non coplanar An y vector n 1 m ma b e e xpressed in terms of them as Th e ’ ’ , y . l ut a nd o -m l l =m ’ m o i m e u Hence ’ n o o n ’ n ’ m xa + yb + ’ = 1, n z c . g ving to r a suitab le value the vector equal to an vector in space By B ut r , , , B ' y r may be made O . O This shows that O r must b e coplanar with d and e Hence O r can take on o nl y those vecto r values which lie in the plane of d and e Thus the assumption that l m n are non coplanar lea ds to a contradi ction Hence I m 11 must be coplanar and the theorem s proved Th e o re m : If a d ad c O be e x pre ss ed as th e sum of thr ee terms . o , , . i y i i y i , . , . of wh ch the ante cedents a b c are known to be non-coplan a r the d ad c O can be reduced to a sin gle dyad when and onl when the consequents 1 m n are collinear Th e proof of the fi rst part was given above To prove the second a t supp o se O could be e xpresse a s , , y , , , . pr d O z dp . T= a l x p + b m x p + c n x u . VECTOR A NAL YSIS 2 84 From the second equation it postfactor for any vecto r where a ’ , ’ b c , t From ’ is evident that is the reciprocal system to a , b, 0 T us e d a i g ve s o the fi rst e xpression Hence y l must be z ero for every value of r tha t is for ever va ue of Hence y z , , , cc , . l x p z m m m s n X p = u Hence I m and n are all parallel to p and t h e theorem has been demonstrated If the three consequents 1 m 11 had been known to b e non coplanar in stead of t h e three antecedents the statement of the theorems would have to b e altered by interchanging the words a n t e c e de n t and c o n s e qu e n t throughout There is a fu r ther theorem dealing with the case in whi ch both antecedents and consequents of O are coplanar Then O is reducible to the sum of two d ads D e n i tio n : A dyadic which cannot be reduced t o the sum of fewer than three dyads is said to be c o mp le te A dyadi c which may be reduced to the sum of t wo dyads but cannot b e reduced to a single dyad is said to b e p la n a r In case the plane of the antecedents and the plane of the con sequents coincide when the dyadic is e xpressed as the sum of two dyads t h e dyadi c is said to be u n ip la n a r A dyadic which may b e reduced t o a single dyad is Said to be lin e a r In case the antecedent and cons e uent of tha t dyad a re col , , . , , , . fi y . . . , . , . q . VECTOR ANA L YSIS 286 y i a d c O us ed as a d f prefactor reduces any ve c tor r to the l ine of the an tecedent of O In particular any vectors perpen di cular to the con sequent o f O are reduced to z ero Th e dyadic O used as a postfactor reduces any vector r to the li ne of the consequent of O In particular any vectors perpen di cular to the ante cedent of O are thus reduced to z ero If O is a z e r o dy a dic th e ve c to rs 8 a n d t a re bo th z e ro n o m a tte r wh a t th e va lu e of r m a y be D e n i tio n : A p la n a r dyadic is said to possess on e de gree of A li n e a r dyadic is said to possess t wo degree s of n u lli ty A z ero dyadi c is said to possess th re e degre e s of n u l n u l l i ty l i ty or c o m p le te n u lli ty Th e o re m : Th e di rect product of t wo complete dya dics is complete ; of a complete dyadi c and a planar dyadi c planar ; of a complete dyadic and a lin e ar dyadic linear Th e o re m : Th e product of two planar dyadics is planar e x cept when t h e plane of t h e consequent of the fi rst dya dic in the product is perpendicular to the plane of the a n t e c e dent of the second dyadic In this c a se the product reduces to a linear dyadic — and only in this case O ; bu t se que n t o no va lu e s o the r Th e . . . . . . fi . . . . , , . . . L et O= Q Th e vector T s r a 1 b 1 O : a 2 b 2, T o . takes on all value s in the and 0 2 a; 0 vector s ’ 8 ’ = O O o 3 1 y c z . takes on the values s 9 4 + { cc (b 2 ° cl ) plane of 0 1 L INEAR VECTOR FUNCTI ONS L et :c where and 9 , ’ a a: (b l 0 a: “2 (3 1 y 1 a 2, y (b l ) 1) ’ 2 87 ’ Ff ! (b z - These equations may always b e solved for a: and y when ’ any desired values cc and y are given — that is when s has any des ired value in the plane of a 1 and e , — unless the determinant ’ ' , . B y b ut bl ° b2 ° Chap II , x b ( . ° 1 b2 ° 2 2 this is merely the product . . 1 bl x b .) <c l 0 > < c .) . vector b 1 b 2 is perpendicular to the plane of the con sequents of O ; and 01 02 to the plane of t h e antecedent s of T Their scalar product va ni shes when and only when t h e vectors are perpendicular that is when the planes are per C ons equently 8 may take on any value in the p e n dic u l a r plane of a 1 and a 2 and O T is therefore a planar dyadi c unless the planes of b 1 and b 2 01 and c 2 are perpendicular If ho w ever b 1 and b 2 01 and c 2 are perpendi cular 8 can take on only values in a certain line of the plane of a 1 and and hence O T is linear Th e theorem is therefore proved T he o re m : Th e product of t wo linear dyadi cs is l inear e x cep t when the cons equent of the fi rst factor is p e rp e n In thi s case the dic u l a r to the antecedent of the second product is z ero — and only in this case The o re m : Th e product of a planar dyadic into a linear is linear e x cept when the plane of the consequents of t h e planar dyadi c is perpendicular to the ant ecedent of t h e linear dyadic In this case the product is z ero — and only in t his case Th e o re m : Th e product of a linear dyadic into a planar dyadic is linear e xcept when the consequent of the linear Th e x , . , ’ . , . ’ ’ . . . . . . VECTOR ANA L YSIS 2 88 dyadic is perpendicular to the plane of the antecedents of the planar dyadic In th s case the pr o duct s z ero — and only in this case It is immedia t ely e vident t hat in the cases mentioned the produc ts do reduce to z ero It is not quite so apparent that Th e proofs are t hey can reduce to z ero in only those cases si m ilar to t h e one given above in the case of two planar dyadics They are left to the reader Th e proo f of t h e fi rst theorem stated page 2 86 is al so left to the reader i . i . . . . . , , Th e I de mfa c tor ; fi 1 . Re c ip ro c a ls f Dy a dic s Conj u a te s g a nd o dyadic applied as a prefactor or as a postfactor t o any vector al ways yields that vector the dyadic s said to be an ide mfa c to r That is De n i t io n : If a i . if O r or if r O r for all values of r r , for all values of r , y then O is an idemfactor Th e capital I is us ed as the s m bol for an idemfacto r Th e idemfactor is a c o mp le te dya di c Fo r t here can b e no di rection in which I r vanishes Th e o re m When e xpressed in nonion form the idemfactor is . . . o . 3 3 ( ) Hence all ide m fa c to rs are equal To prove that the idemfac t or takes t h e form ( 3 3 ) it is merely necessary to apply the idemfactor I to the vecto s i 1: respectively et . j , , . I= L a “ a 1 g In t h e t h e o ry d in a ry a l e bra or . of r u ii + a 21 i j+ a 31 ki + dya d ic s Th e n o ta the i t on m lz i j+ j 22 j a 82 a ls “ kj gg ik k j 23 a as kk i d e fa c t o r I pla ys a mi le is in t e n de d t o s u e s t t h is . l a na o a na l o gg o us y . to un i t y in VECTOR ANAL YSI S 290 l the fi rst place since O is t h e idemfactor it is a comp ete dyadic Hence the antecedents a b c are non-coplanar an possess a set of reciprocals a b c et In , , . ’ ' By y r h pothesis Then ’ , r , . O L r. r for all values of r that is for a ll values of a y 2 correspondi ng coe ffi cients must be equal That is , , , , . m =b y ’ n , Hence , . c z ’ . i d an two dyadics and f the pro uct 1 O T is e ual to the idemfactor ; then the product T O when the facto s a re taken in the revers e d order is a so e ual to the idemfactor Th eo re m q q : IfO and T be d r l , , . L et To , O show T Tz o o I . O= I . -( O T) = r -I = r = O T rO O r ) ( r , , Thi s relation holds for all values of r A s O is complete must take on all desired values Hence b y de fi nition . r O . T o O= I . If the product of two dyadics is an idemfactor that pro duct may be taken in either order De n itio n : When two d adics are so related that their product is equal to the idemfactor they are said to b e , fi . y , 1 of i Th is m t wo inc o m d e f a c to r . m m it a t e s bo t h t h e dya di c s 4» a n d 11! t o b e c o pl e t e Fo r t h e pro d u c t ple te dya dic s is in c o ple te a n d h e n c e c o uld n o t b e e qua l t o th e n ec e ss . L INEA R VECTOR FUNCTI ONS 291 l nota t ion used for reciproca s in ordinary alge b ra is emplo ed to denote reciprocal dyadics That is re c ip roc a ls 1 . Th e y . I —1 T Th e o re m : T a nd O , I _1 w O Reciprocals of the same or equal dyadics are equal et O and T b e two given equal dyadics O their reciprocals as de fi n ed above B y h pothes s L . y . a 42 os — — 1 — -1 -1 — 0 O ( . a O T, : - 1 . O . O I l — O ‘ Hence O : 1 . I ‘ l — O 1 T 1 . T : O . —l O —1 2 . O O . -l T . T . , -l , I, T I ‘ T I “ d 1, «H ‘ O an ‘ -1 7 9 -l o i I T, : n , " I = T “ l 4 . reciprocal of O is the dya di c whose antecedents are the eciprocal system to the cons equents of O and whose c o n s e quents are the reciprocal syste m to the antecedents of O If a complete dyadi c O b e written in the form r Th e . O its reciprocal is O ‘ bm : I l ’ a ’ ’ m b o n, ’ ’ n c ’ 6 3 ( ) . Fo r the direc t products of a complete dyadic d adics T and 9 are equal as dya di cs then T and 1 A i pl t dy di h l ( fi i t ) ip Th e o re m into two y : If n nc o m e e a c a s no n e re c ro c a . O Q VECTOR A NAL YSIS 292 r are equal If the product of a dyadic O into two vecto s and 3 (whether the m ul tiplication b e pe rformed with a dot or a cross) are equal then the vectors r and s are e ual That is r - . q , . , if O a nd if a nd if This o O Q, T= O- then T = O, = O then r a r s, o the n y y = s, s. r y ma be seen b multiplying each of through b the reciprocal of O , T ‘ l O O ' T o 1 ' T o -1 T T: : 0 T g a -O 1 q reduce the last e uation proceed as follows an vector y I I r = t s t To x , y x o t Henc e q e uation s the I o . Le t t b e , 17 . r is an vector is equal to 8 Equations ( 3 7 ) give what is e uiva ent to the aw of celation for complete dyad cs Complete d adi cs ma b e canceled from either end of an e x pression just as if the were sc a lar quantities Th e cancelation of an incomplete dyadic is not admissib le It corresponds to the cancelat on of a z ero factor in ordi nar algebra Th e o re m Th e reciprocal of the product of an numb er of d a di cs is equal to the product of the rec proca s taken in the op p o s ite order It will be s u f fi cient to give the proof for the c as e in wh ch t h e ro uct c ons s ts of t wo d a d c s To s how As t , q l . i . y l y i . . y y l . y i i . p d y i y i . VECTOR ANAL YSIS 294 G eometricall y the transformation r = Or ’ r is a reflection of space in t h e jk-plane This t ans formati on replaces each fi gure b y a symmetrical fi gure symmetrically si t uated upon the opposite s ide of the k-plane Th e trans formation is sometimes called p e rve rs io n Th e idemfac t or has a l so a doub l infin ite system of square roots of the form . j y . . T G eometricall , jj ii kk . y the transformation r ’ = Tr is a reflec t ion in the i ax is This transformation replaces each fi gure b y its equal rotated ab out t h e i-a x is through an angle of Th e idemfactor thus possesses not only two square roots ; b u t in addition two doubly in fi nite systems of square roots ; and it wil be seen ( Art 1 2 9 ) that these are b y no means all Th e c o nj u a te of a dyadic has been de fi ned A 9 9 r t g ( ) as the dyadic ob tained by interchanging the antecedents and consequents of a given dyadic and the nota t ion of a sub script C has b een employed Th e equation - . l . . . ’ . h as b een demonstrat ed Th e following theorems concerning conjugates are useful Th e o re m : Th e conjugate of the s u m or di fference of t wo dyadics is equal to the sum or difference of the conjugates . . , T ¢ : t ( ) C = w e : l: T0 . conjugate of a product of dya di cs is equal to the prod uct of the conjugates taken in the op p o s i te or er The o re m : Th e d . L INEAR VECTOR FUNCTI ONS ill 295 w be s u fli c ie n t to demonstrate t h e theorem the product contains two factors To show It in case . m 0 n (i r r H ence O ) ( wc wm : ' -( T -O Oo z T = TC (r r o r, -O) = TO T) O = TC O ( -a r . O0 . . y i conjugate of the power of a d adic s the power of the conjugate of the dyadi c Th e or e m Th e . 2b y Thi s is a corollar of the foregoing theorem Th e e xpression O 2 may b e interpreted in either of t wo equal ways Th e o re m : Th e conjugate of the reciprocal of a dyadi c is equal to the reciprocal of the conjugate of the dyadi c . . . . 1 7 1 ( 6 Fo r Th e the (T A R A To = ( T Ia o I : . y idemfactor is its own conjugate as ma be seen from nonion form . I=i i + ji + k k O0 H enc e ( ¢ ) C —l $0 : W ) ( I C 0 To . Hence expression O0 may therefore be interprete d in either of two equivalen t ways — as the reciprocal of the conjugate or as the conjugate of the reciprocal D e n i tio n : If a dyadic is equal to its conjugate it is said If it is equ al to the negative of its con uga te to be s e lf-c o nj Th e " fi 1 . , . 2 96 jugate it dyadi cs is , s VECTOR A NAL YSIS i a d to be a n ti se fc o n u ga te j l Fo r . fconjugate se l - . Fo r anti-self-conjugate dyadics r -O O — O r, o OO — . dyadic may b e divided in one and way into two parts of which one IS se f-conjugate and the other anti-self-conjugate An y Th e o r e m l . Oa ) — . $00 ¢ 0 ¢a + $2 — Q Hence the part %( O O0) is sel f-conjugate ; and the -se f-conjugate O O anti Thu s the div sion has bee n 0) H accompl shed in one way et i l , L . i . (O ¢ a) O ( To ) T T ’ W O T ” . ” . Suppose y it were possible to decompose O in an other wa into a self-conjugate and an anti -self-conjugate p a rt et then L . T Whe e ( r O Hence if ( O T ( " ’ (T ’ : O) ’ Q) Q) Hence if ( O conjugate ” ( O + 9 ) ( ’ 9 is self-conjugate T ( H Q) T ” Q , Q)0 = $ is self-conjugate " is anti -self-conjugate . Q : 9 0, Q 9 0. . QU QO ‘ ‘ o 0 Q ‘ is an ti -self VECTOR A NAL YSIS 2 98 §( O — Oc ) o O r r x r, X ¢ x o conjugate dyadi c O possesses one degree of nullity It is a uniplanar dyadic the plane of ” whose consequents and antecedents is perpendicular to O the vector of O This theorem follows as a corollar from equa t ions Th e o re m : A n y dyadic O may b e b roken up into two parts of which one is self-conjugate and the other equivalent to minus one half the vector of O used in cross multiplication Th e o re m : An y ” a n t is e l f. x , y . . l 1 or symbolically TX X 2 . ( 45) . vector 0 used in vector m ul ti linear vector function Fo r An y plication de nes a fi . Hence it must be possible to represent the Operator c x as a dyadic Thi s dya dic will be uniplanar with plane of its antecedents and consequents perpendicular to 0 so that it will reduce all vectors parallel to c to z ero Th e d adic ma be found as foll ows . , . Ry 3 1 ( ) (I x ) o , {(I x r I Hence and -I I c -( c x = I r ) } {I r I) c I r x ( ) . -( I x c ) This may be sta ted in words x c r c -r r y r . ( c x I) r , y L INEAR VECTOR FUNCTI ONS 29 9 vector c used in vector multiplication with a vector r is equal to the dyadic I x c or c I used in direct mul tiplication with r If e precedes r the dyadics are to be us ed as p re f a c t o rs ; if c foll ows r as dyadics o f a r s t c t o s T h e p I x c and c I are an t i -self-conjugate In case the vector 0 is a unit vecto r the application of the operator 0 to any vector r in a plane perpendicular to c is equivalent to turning r t hrough a posi t ive righ t angle ab out the ax is 0 Th e dyadi c c x I or I c where c is a unit vector therefore turns any vector r perpendi cular t o 0 t hrough a right angle ab out the line 0 as an ax is If r were a vector lying out of a plane perpendicular to c the e fle c t of t h e dya dic I woul d b e to annihilate that component of r whi ch I x c or c is parallel to c and turn that component of r whi ch is p e rp e n dic u l a r to c through a right angle a b out 0 as a x is If the dyadi c b e applied t wice the vecto rs perpendicular to r are rotate d through t wo right angles They are reversed in direction If it be appli ed three times they are turned thr ough three right angles A pplying the operator I c or c I four times b rings a ve c tor perpendicular to 0 b ack t o its original osition Th e powers of the dyadic are therefore Th e o re m : Th e x . x x , . . x . . x . . x . . p x . — — I ( — c c ) , I > <c — c c c x , x I . thus appears that the dya di c I c or c x I ob eys the same law as fa r as its powers are concerned as the scalar imaginary 1 in alge b ra V I is a quadrantal vers or only for Th e dyadic I c or c vectors perpendicular to c Fo r vectors parallel to c it acts as an annihilator To avoid this c fi e c t and obtain a true It ‘ x . x . . VE C TOR A NA L YSIS 8 00 quadrantal versor for all vectors r in space it is mere sary to add the dyad c c to the dyadic I x c or c x I ly ne ces . X — I, X — 2 3 X, X = I, 4 X =X 5 i . dyad c X therefore appears as a fourth root of the idemfactor Th e quadrantal versor X is analogo u s to t h e imaginary V 1 of a scalar algeb ra Th e dyadic X is com s ts of two parts of whi ch I x c and consi s anti self l e t e p conjugate ; and c c self-c onjugate Ifi k are thr ee perpendicular unit vectors Th e . i . , j , . , I > < i i x I = k j— jk — k i, =k > i — ij <I = j , k y , as ma be seen by multiplying the idemfacto j I ii ii r k 1: into i and k successively These e x pressions represent quadran t al ve rs o rs ab out the ax is i k respectivel comb ined with annihi lators along those ax es They are equivalent when u sed in direct multiplication to i k x respectively , . , , j y , . , — x , , , i (j — ij ), e pression is an idemfactor for the plane of i and b u t an annihilator for the direction k In a similar man ner the dyad k k is an idemfactor for the direction k but an Th e j , x 4 k (Ix ) . , VE C TOR A NAL YSIS 302 115 L et ] vector f Dy a dic s Re du c t ion t o N o r ma l For m o y i any complete d ad c and let Then t h e vector r O be r be i un t a ’ . ’ r = Or ll i is a linear function of r When r takes on a v a lues co us s tent wi t h its b eing a un i t vector that is when the termin us of r describ es the surface of a un it sphere — the vector r varies continuous l and its terminus describes a surface This 1 surface is closed It is fact an ellipsoid Th e o re m It is always possi b le to reduce a complete d a di c to a sum of thr ee term s of which the antecedent s among themselves and the cons equents among themselves are mutu ally perpendic u lar This is called the n o rm a l fo rm of O . , ’ y , . y . . . . O a ’ ’ b jj i i ’ k k c . r d i d demonstrate the theore m consider the su fac e escr be To r ’ = Or . this is a closed surface there must be some directi on of r ’ which makes r a max imum or at any rate gives as great a value as it is possib le for r to ta k e on et this d rection of r b e called 1 and let the correspondin g direction of r ’ the direction in whi ch takes on a value at e a st as great as any — b e called a Consider ne xt all the values of r which li e in a plane perpendicular to i Th e corresponding values of I lie in a plane owing to a fact that O is a linear vector As ’ L ’ . r , r i ’ l . r . ’ 1 Th is m a y b e p ro ve d a s fo ll o ws ’ r z H e nc e g r . dn non on d d es c rib e s re e is t h e . e He n c e llipso id . ’ r mq fo r B y e x p re s s in g ‘ lfin e -1 r —l w f ( Tc r=1 i r -1 = ¢ r a , ' o qua t io n r ua d ri c s urfa c e th e e ' a . ? ’ r Th e z ] is on se e n . t o b e o ft h e se c o n ly c l o se d q ua dric f d s ur a c e L INEAR VECTOR FUNCTI ONS O f these 3 03 function values of r one mus t be at least as great as any other Call this b and let t h e correspondi ng directi on Finally choose k perpendicular to i and of r b e called upon t h e positive side of plane of i and et 0 b e the value of r whi ch corresponds to r k Since the dyadic O changes i k into a b c it may b e e xpressed in the form . j . ' j j L . . ’ , j . , , , O i a b j k c . rema ins to show that the vectors above are mutually perpendicular It b, a, as determin ed c . r ’ o dr ’ ’ r o i a i -d r + b j+ c k) + b j+ c k) + r ’ r , d r, o b jdr o r + ’ o c k -d r . When r s parallel to i r is a maximum and hence must be ’ i , y perpen dicular to d r Since r is a unit vecto r d r s alwa s perpendicular to r Hence when r is parallel to i ’ . . r ’ o b j dr + r ’ c -d r = k j 0 . further d r is perpendic ul ar to r c vanishes and if d r is perpen di c ul ar to k r b vanishes Hence when r is parallel to i r is perpen di cular to b oth b and 0 B u t when r is parallel to i r is parallel to a Hence a is perpendi cular to b and 0 Consider nex t the plane of and k and the plane of b a n d c et r b e any vector in the plane o fjand k If ’ , ’ , o c , . ’ . , ’ . . L . ( b j+ r ’ o j . , dr ’ r ’ When r takes the value di hence is perpen cular b o jr , to c j dr k) + r r , ’ c c k is a max imum in this plane and Since r is a un t vector it is dr ’ ’ . i VECTOR 3 04 r A NA L YSIS r is perpendi cul ar to d Hence when is perpendic ul ar to and j . ll l j para e to , dr , k -d r . j Hence r c is z ero B u t when r is parallel to r ta kes t h e value b Consequentl b is perpen di cular to 0 It has therefore been shown that a is perpendicular to b and C onsequentl the three c and that b is perpendicular to 0 antecedents of O are mutu ally perpendic ul ar They ma be ’ ’ denoted b i Then the dyadi c O takes t h e form k ’ y . . ’ , . y . , y j y . ’ , . , O= where a ’ ’ b jj+ i i c ’ k k, are scalar consta nts positive or negative Th e o re m : Th e complete d adic O may alwa s be reduced to a sum of three dyads whos e antecedent s and whose cons equents form a right-handed rectangul ar system of unit vectors and whose scalar c o e fli c ie n t s a e either a positive or a ll negative a , b, y c . y r ll . O : z iz ( a 5 3 ( ) ’ ’ i i + b jj proof of the theorem depends upon the statements made on page 2 0 that if one or thr ee vecto s of a righ t -handed system be reversed the resulti ng s stem is left-hande d but if two be reversed t h e system remains right-handed If then o n e of the c o e fii c ie n t s in 5 2 s negative the d rections of the ( ) other two ax es may be reversed Then all the coeffi cients are negative If two of the coe ffi cient s in ( 5 2) are negative the directions of the two vectors to which they belong ma be reversed and then the coeffi cien t s in O are all positive Hence in any case the reduction to the form in which all the coe ffi cients are positive or a are negative has been performed As a l imitin g c ase between that in which the coe ffi cients a re a o s it v e a n d t ha t n wh ch the a re a ll nega t v e comes Th e r y , i , i . . y . , . ll . ll p i i i y i VECTOR A NAL YSIS 3 06 O0 O O, z a : O ’ i i a z s ’ ii O O a 2 I (b ( O z a z I ( j c s j z c 2 O a z — O z 2 a a Z — O a ’ z k k kk ’ , . O0, ’ ’ = i i + kk b ( : ’ O, = . I = ii + j j O2 ’ ’ ’ ) jj z 2 z l) . ’ ’ jj 2 (c ’ k k, a 2 )k ’ k ’ , i = 0 ’ o ) jj l) 2 o (o — z z d ) k k, i = 0 . i were not parallel ( O 2 a I) wo ul d annih late two vectors i and i and hence ever vector in their plane 2 2 a I) would therefore possess two degrees of nullit (O and b e linear B u t it is apparent that f a b c are di fle re n t this dyadi c is not linear It is planar Hence i and i must be parallel In like manner it may b e shown that and k and k are parallel Th e dyadic O therefore takes the form If i a nd i ’ 2 y ’ y . i . . , , j . . ’ j ’ , ’ . O= where a , b, a i i + bj j+ c kk are pos itive or negative scalar constants c fi Do u ble Mu ltip lic a tio n . 1 product of two dyads is t h e s c a la r quantit obtained by multipl ing the scalar product of the antecedents by t h e scalar product of the consequents Th e product is denoted b y inserting two dots between the De Th e do u ble do t n i t io n : y y . dya ds “ a This Th e p rin te d y : c d= a -c b d ( 56 ) . product evidentl obeys the commutative law a 1 b re s e a rc h e s fo r t h e fi rs t m of t i b z c dz Pro fe ss o r Gibbs e . c d u po n z a D b, o u bl e M ult ip lic a t ion a re h e re L INEAR VECTOR FUNCTI ONS 3 07 and the distributive law both with regard to the dyads and with regard to the vectors in the dyads Th e doub le dot produc t of two dyadics is ob tain ed b y multiplyin g the prod u c t out formally accor di ng to the distri b utive law into the sum of a number of double dot products of dyads . . T= a c 2 1 d1 + h2 c 2 d2 + 03 d3 + b 3 3 a 3 ( ° 1 d1 ez aa c 3 _ a 3 1 b1 :c l d1 a l bl z c 2 d2 + a 2 b2 261 d1 + a 2 b2 3 02 d2 + d1 a a bgz d2 a gba z c l c z b3 fi bi i c a ds i a , b, z c a d3 + a ba i cs ds l a ' d3 + ' 0 d2 + a 3 3 product of t wo dyads is the dy a d of which the antecedent is the vector product of the ant ecedents of t h e two dyads and of which the consequent is the vector product of the consequent of the two dyads Th e product is denoted b y inserting t wo crosses between the d ads De n i ti o n : Th e do u ble c ro ss . y b x d . This l y i 5 ( 7) product a so evidentl ob eys the commutat ve law VE C TOR ANAL YSIS 3 08 d d d and the distrib utive law both with regard to the dyads an wi t h regard to t h e vectors of which the dyads are compose Th e doub le cross product of two dya di cs is therefore de fi n e as the formal e xpansion of the product accordi ng to the distrib utive law into a sum of double cross products of dyads . . If and i ( 01 d1 + c 2 +c + a 3 b3 § c l d1 + u b l x d1 + a 1 x + + a > < 03 11 3 > < d1 + 8 c2 b z x d3 + b2 > < d2 > <c l d3 + - a 3 a2 3 d2 a 3 m > < c 2 b3 > < d1 double dot and doub le cross products of two dyadics obey the commutative and distrib utive laws of multiplication B u t the double products of more than two dyadics ( whenever they have an meaning) do not obe t h e associative l a w Th e o re m : Th e y . y . O O ( O z T T O = T T§ O § m ww y z wm i . theorem is su ffi cientl evident w thout demonstration . VECTOR A NAL YSIS 31 0 the factors is reversed each scalar triple product cha nges sign Their product therefore is n o t a l t e re d A dya di c O may be multipl ied by itself w th doubl e cross et i i . . L . O= a l + bm + c n l x m + < + b > + > <a c n products in the main equal in pairs Hence Th e l > <n c m x m + b x <b m x l + b > a > < a di agonal x m + vanish > < c . m x c n c Th e x n n . r othe s are . m x If a , b , c and l , m, n n + c are non —coplanar this 2 [ a b l x m) x a c ] l m [ n a l l ( ] , ’ ma y + b m . 60 ( ) be written ’ c r n r ) , product O 35O is a species of power of O It may be re garded as a square of O Th e notation O2 will be employed to represent thi s product after the scalar factor 2 has been stricken out Th e . . 2 (b x c mxn + c l x m) > <a 6 1 ( ) triple product of a dyadi c O e xpressed as the sum of three dyads with itself twice repeated is Th e O O2 : O= (b i Oz O= 3 n O l x m) x c y e xpanding this product ever term in whi ch a letter is repeated vani shes Fo r a scalar triple product of t hree vec In . L INEAR VECTOR FUNCTI ONS i to rs two of wh ch are equal is z ero reduces to three terms only O2 : O= [ b o a ] [ m n l] O2 O : : p [c a b] 3 [a b Hence the product . [ n l m] c 31 1 [ a b c ] l m n [ ] ] [l m n] l m n [ ] y . triple roduct of a d ad ic b y itself twice repeated is equal to s ix times the scalar t riple product of its antecedents multiplied by the scalar triple product of its consequents Th e product is a species of cu be It will b e denoted b y O3 after the scalar factor 6 has been stricken out Th e . . . O i l O g z [ a b c ] I n [ n ] 2 6 ( ) . be ca led the se c o n d of O ; and O3 the third of O the following theorems may be stated concerning t h e seconds and thir ds of conjugates reciprocals and products Th e o re m Th e second of the conj u gate of a dyadic is equal to the conjugate of the second of that dyadic Th e third of the conjugate is equal to the third of the dyadic If O2 , , , , . . . w ( a s : 63 ( ) $3 second and third of the reciprocal of a dyadic are e ual respectively to the reciproca s of the second and third The o re m : Th e q l . — «0 9 ( D « O 2 2 ( O 3 —1 _ l l a 1 fig —1 1 + bm a l l Q2 $3 l m b + [ c ] ] 11 0 l b 1 n l m [ ] 1 -1 o n ’ ’ a . VECTOR ANAL YSIS 31 2 [ a b c [l m n ] ] l ( w )2 —1 B [ ut a Hence ( b c [ ] l b l c ] —l m ’ 1 $8 [ b a ] c n n ’ ] O2 ( O )2 “ l m l [ and [ 1 n o ) n o l ’ 1 O2 ) a mb a mb a l l ( l ] 1 l m n [ ] . 4 . l m n , [ ] 1 [ 1 ( O )3 ‘ Henc ( e Og) [a l " ( a ’ 0 ’ ’ b O ] b ‘ c l n l m [ ] ] [l ’ ’ m O3 )a , " 1 . r q second and thi rd of a product a e e ual respecti vely to the product of the seconds and the product of Th e o re m : Th e t h e t hirds . (O T) 2 = O 2 o 9 Choose 03 -T2 $3 T3 : y , q an three non-coplanar vectors 1 m n as cons e uents of O an let m n be th e ante cedents of T d , ’ , ’ . , O ( e x f+ > < c fx d + a a > <b d x e, l > < m, d x Hence Henc e O2 e ( O o x f+ c > <a T) 3 = [ a b c ] fx d+ d e f [ ] a > <b c . dx e . VE C TOR ANAL YSIS 31 4 O2 = b > <c m x n O3 + c > <a [ b c a ] l > < m, n m l [ ] . necessary and suffi cient con di tion that a dyadic O be complete is that the third of O be d fferent from z ero Fo r it was shown ( Art 1 06) that b oth the antecedents and the consequents of a complete dyadi c are non-c oplanar Hence the two scalar triple produc t s whi ch occur in O3 cannot vanish Th e o re m : Th e necessar and su ffi cient condition that a dyadic O be planar is that the third of O shall vanish but the second of O shall not vanish It was sho wn ( Art 1 06 ) that if a dyadic O be planar its con sequents l m 11 must be planar and conversely if the c o n s e quents be 0 0planar the dyadi c is planar Hence for a planar dya di c O8 must vanish B u t O2 cannot vani sh Since a b e have b een assumed non-coplanar the vectors b x c c a a x b are nonHence if O2 vani shes each of the c oplanar vectors m x n n x l 1 x m vanishes — that is l m n are col linear B u t thi s is impossible since the d adic O is planar and not linear Th e o re m : Th e necessar and suffi cient condition that a non vanishing dyadic be linear is that the second of O and cons equently the third of O vanishes Fo r if O be linear the consequents l m n are collinear Hence their vector products vanish and the consequents of O2 vanish If con versely O2 vanishes each of its conse uents must b e z ero and hence these consequents of O are collinear Th e vanishing of the third unaccompanied by the vanish ing of the second of a dyadic implies one degree of nullity Th e vanish n g of the second implies two deg ees of n ul li t Th e o re m : Th e i . . . y . . . , , . x . . , , , , , . , y , . . , , , y - , , . , . , . , q , . , i , r y . . L INEAR VECTOR FUNC TI ONS i i y lli y van sh ng of the dyadic i t self is complete nu t res ults ma be put in tab ular form Th e 31 5 . Th e . O3 O3 = O, O3 = O, is complete O 0, O2 O2 = O, . 0, O is p la n a r O 0, . O is lin e a r y . follo ws imme di ate ly that the third of an anti -self-conjugate dyadi c vanishes ; b u t the second does not Fo r any such d adic s planar but cann ot be linear It y i . . N on io n Fo rm If O D e te r mi n a n ts . conjugate of they are ar ange d agon al Thus O ul 1 ii + a “2 1 ji “2 2 a 31 ki + i j+ a ll “ 12 “ 32 f n ts o a Dya dic ik 13 k l 2s kj a kk 33 . has the same scalar coeffi cients as O but s mmetricall with res pect to the main r d y i I n va ri a . be ex pressed in nonion form O= Th e 1 y , . O0 : a 11 “ a ii a i j 12 la 21 a 31 i k, 32 1 k, a jj “ k j 23 a “2 2 ki i j+ 33 kk . second of O may be computed Take for instance one term et it b e required to fi n d the coe ffi cient of i jin O2 What terms in O can ield a doub le cross product equal to i Th e vector product of the antecedents must b e i and the vector product of the cons equents must b e Hence the antecedents must be and k ; and the conse uents k a n d i These ter ms a re Th e . L y j . j q “as o “3 1 1 m Th e re s u l ts de t e r in a n ts , , . g h o ld k1 x kk o mg “2 1 “3 34 “ 3 1 “2 3 0 . . , 1 0 ” g l y fo r de t e r i n a n t s o ft h e t h i rd o rde r e bra A t e u l h M h r o u t i s i d ipl l r r s o e h r e h of on j . . Th e e xt i e ns o n to 31 6 j VECTOR A NAL YSIS Hence the term in i in O2 is “ 3 2 Th is is the rst minor of fi i Th s i l l sa “ “ m the determi na n t 21 a i i minor is tak en w th the negative sign That s the coe ffi cient of i in O2 is what is termed the c ofa c to r of the coeffi cient of i in the determinant Th e cofa cto r is merely the fi rst minor taken with the positive or negative sign accordin g as the sum of the subscripts of the term whose fi rst minor is under consideration is even or odd Th e c o fi c ie n t of any dyad in O2 is easily seen t o be the cofactor of ef the correspon di ng term in O Th e cofacto s a re denote d generally by large letters j j . , . . r . . a a i r rs the cofacto of a is the cofactor of a is the cofactor of “ “ a 21 i With th s notat on the second of O becomes A 2 1 ji + A 2 2 Aal k i Th e value of the third of as the sum of thre e dya ds O “ 21 j “ 31 “4 “ . A 32 k j O Az s k k A3 3 k k . y ri i may be obta ined b w t ng 19 1 4 " 22 l “3 2 k“ 2a j “s a l0k “ “ O VECTOR ANALYSIS 31 8 Ifthe determinant be denote d by D A2 1 A 11 _ If T is A31 “ 77 D A1 2 A2 2 D D A1 3 A2 3 1 1) 0 A 82 ' k ° ’k D 88 y a second d adic given in nonion form T = bu i i + bu i j as bl s i k, ”2 3 311 , ba s k k , b,, k j ba l k i y dily be fo nd the product O T of the two dyadics ma rea by actuall performin g the mul tiplication y T T ”1 1 “ “ 13 1 “ “ 22 “ 23 ” “3 3 ”1 3 ”2 1 “2 3 ”3 2 ) j j ”1 1 1 “ “ 32 ”2 1 “ 33 11 ”1 1 ”2 3 “2 ”1 2 2 ”2 3 “ 12 “ 3 2 ”2 3 33 ”1 2 “1 3 ”2 1 “ 22 ”2 2 “2 3 ”2 3 “3 ”3 1 “3 2 ”3 2 “ ”3 3 Since 33 ”3 3 2 ”2 2 >k 11 ”1 3 “2 1 1 ”2 2 ”3 3” k “3 “ ”2 2 >1 k “2 2 ”1 2 “ 32 ”3 3 13 ”3 011 1 4 “ “1 2 “ ”3 1 > l1 ”1 2 “ ”1 2 “1 2 ”1 3 ”3 2 ) k l $1 ? ”3 1 ) 1 1 “1 3 ”3 2 ) 1 j ”1 1 ” 12 ”2 1 u : the third or determinant of a product is equal to the product of the determinants the law of multiplication of de t erminants follows from ( 6 5) and , L INEAR VECTOR FUNCTI ONS “ ”1 1 “ ”2 1 “ ”3 1 “ ”1 1 “ ”2 1 “ ”3 1 “ ”1 1 “ ”2 1 “3 ”3 1 11 21 31 “ 11 ”1 2 “ ”22 “ “ 21 ”1 2 “ 22 ”22 “3 1 ”1 2 “ 32 ”2 2 12 31 9 12 22 32 13 23 13 ”3 2 “ 11 ”1 3 1 “ 12 ”2 3 “ 13 ”3 3 : “ 23 ”3 2 “ 21 ”1 3 “ 2 2 ”2 3 “ 3 2 ”3 3 , “ 33 ”3 2 “ ”1 3 “ ”2 3 “ ”3 3 31 " “ 32 33 3 6 7 ( ) ° rule may be stated in words To multiply two deter minants form t h e determi nant of which the element in the wit h ro w and n t h column is t h e sum of the produc t s of t h e elements in the m t h row of the fi rst determinant and n t h column of the secon d Th e . . If O2 = b x l > <m e . Then c Hence Hence I O2 ! IO 2 ( x a [ 2 1 93 A A A A A A A A A 33 a a b x b] n m x [ l m n] [ 2 11 O3 “ . determinant of the cofactors of a given determinant of t h e third order is equal to the square of the given determinant 1 22 ] that is A dyadic O has thr ee scalar invarian t s three scalar quanti t ies which are independent of the form in which O is e x press e d These are Th e . . T3 , ”a , the scalar of O the scalar of t h e second of O and the thi rd or determinant of O If O b e e xpressed in nonion form these quantities are , , . VE C TOR 3 20 $8 A NAL YSIS 11 23 A 33 A 22 A1 1 3 “ “2 2 “ “ 13 12 “ 23 22 “ “ 2 3 33 r No matter in terms of what right-ha nded ectangular system of these unit vectors O may b e e xpress ed these quantities are the same Th e scalar of O is the sum of the three coeffi cients in the main di agonal Th e scalar of the second of O is t h e sum of the fi rst minors or cofactors of the term s in t h e main diagonal Th e third of O is the de t erminant of the coe ffi cient s These three invariants are b far the mos t important that a dyadi c O possesses The o re m : A n y dyadic satis fi es a cub ic equa t ion of whi ch t h e three invariants OS O2 3 O8 are the coeffi cients . . y . . B y (6 8) ( . , , O ( O A DC O ( : “ “ 3 “ 12 13 — 27 “ 22 23 “ “ “ Hence ( O— c )3 = Os — cc Ow 9” 33 32 -a — l ' 2 — az 3 as may b e seen by actually p e rfo rm in g t h e e xpansion ' ( O — T — I) 2 O ( x l)0 = Os — x . — a s S O2 5 + s This equation is an identi t y holding for a ll values of the scalar ac It therefore holds if in place of th e scalar a the dyadic O which depends upon nine scalars be sub stituted That is 3 2 O = O l )a ( O O I)3 O I O3 O Oz s i O OS ( ' . , , . o B ut . the te rms upon the left are identically z ero $ 2 $2 8 O $8 1 = O ° . Hence 3 22 VECTOR A NAL YSIS p dyadics are equal when they are equal as o erato rs upon all vectors or upon three non-copl a nar vec to rs That is when T r for all values or for three non O r coplanar values of r r T for all values or for three non O r coplanar values of r T r for all values or for thre e non O r s s coplanar values of r and 3 Two . , , , . y A n y li near y i vector function ma be represented b a dyad c Dyads obey the dis t ributi ve law of multipl ication w th regard to the two vectors composing the dyad + i . l + c O 0 i Multiplication by a scalar is associati ve In v rt u e of these two laws a dyadic may be ex panded into a sum of nine terms by mean s of the fundamental dyads . , O= a _ a u i i, ij i k, , 13 ii k i, k ii + ki + a a 1k j , . k k, m i j+ 32 a k j+ 13 a i lg 33 kk . If two dyadi cs are equal the corresponding coe ffi cients in the i r expans ions into nonion form are equal and conversely . L INEA R VECTOR F UNCTI ONS 3 23 dya dic may b e e x pressed as the sum of three dyads of whi c h the an t ecedents or the consequen t s are any t hree given non -coplanar vectors This e x pression of the dyadic is uni que Th e symbolic product a b kn own as a dyad is the most general produc t of t wo vectors in whi ch multiplication b y a scalar is associative It is called t h e indeterminate product Th e product imposes fi ve condi t ions upon t h e vectors a and Their directions and the product of their lengths are b determined by the product Th e scalar and vector products are functions of the indeterminate product A scalar and a vector may be obtained from any dyad c b y inserting a dot and a cross b etween the vectors in each d ad This scalar and vector are functions of t h e dya di c An y . . . . . . i . y . . direct product of two dyads is the dyad whose ante cedent and cons equent are respectively the ant ecedent of the b t rst dyad and the consequen of the second multiplied fi y the scalar produc t of the conse uent of the fi rst dyad and the antecedent of the second Th e q . ( a b) 0 ( e d) (b o c ) direct product of t wo dyadics is the formal e x pansion i c ord ng to d stri utive law of the product nto the b i t h e a c Th e i , , VECTOR A NAL YSIS 3 24 sum of products of dyads Direct multi p lication of dyadics or of dyadics and a vector at ei t her end or at both ends ob e s the distri b utive and associative laws of multiplicat ion Co n sequently such e xpressions as y . . w wf 0 0 SO , Q ‘ wr ‘ , y written without parentheses ; for parentheses ma b e ins erted at pleasure without altering the value of the product In case the vector occurs at other positions than at the end t h e product is no longer associative Th e skew product of a dyad and a vector ma be de fi ned b the equation r a b r b x a X ( ) ma y b e . y . y , 2 8 ( ) i q skew product of a dyadic and a vector s e ual to the formal expansion of that product into a sum of products of dyads and that vector Th e statement made concerning t h e associative law for direct products holds when the vector is connected with the d adi cs in skew multiplication Th e e xpressions Th e . y -T e , O o T x r, . e o i s, r -O x s, e x s ( 29 ) y may b e written w thout parentheses and parentheses ma b e inserted at pleasure without altering the value of the product Moreover -O . , O o ( r T x . the parentheses cannot be omitted Th e necessar and suffi cient condition that a d ad c ma be reduced to the sum of two dyads or to a single dyad or to z ero is that when e x pressed as the sum of three dyads of which the antecedents ( or consequents ) are known B ut y . , y i y VECTOR ANA L YSIS 326 Ifthe product of t wo complete dyadics is equal to the idem factor the dyadics are commutative and either is called the reciprocal of the other A complete dyadic may be canceled from either end of a product of dyadics and vectors as in ordinar algeb ra ; for the cancelation is equivalent to mul t iplication b y the reciprocal of that dyadic Incomplete dyadi cs pos s ess no recipro c als They correspond to z ero in ordinary algeb ra Th e reciprocal of a product is equal to the product of the reciprocals t aken in inverse order . y . . . . O ( . 51 —1 - T 0 l O —l ( 3 8) . conjugate of a dyadic is the dya di c o b tained b y in ter changing the order of the an t ecedents and consequents Th e conj u gat e of a product is equal to the product of t h e con u a t e s taken in the Opposite order jg Th e . . 4 0 ( ) conjugate of the re c iprocal is equal to the reciprocal of the conjugate A dyadi c may b e divided in one and o nl y one way into the sum of two parts of which one is self conjugate and the other anti self—conjugate Th e . - . — a > 3 4 ( ) . dyadi c or t he anti -self-c onjugate part of any dyadic used in direct multipli cation is equivalent to minus one-half the vector of tha t dyadi c used in skew m u ltiplication O x r r An y i ant self-conjugate , , , )( r $0) A I X $10 4 4 ( ) dyadi c of the form c x I or I x c is anti-self-conjugate and used in direct multiplication is equiva ent to the vector 0 u sed in s kew m ul tiplication l . L INEA R VECTOR FUNCTI ON S ( c x I) 3 27 r, a -O . x dya di c c x I or I c where c is a unit vec t or is a quad rantal versor for vec t o rs perpendi cular to c and an annih ilato r for vectors parallel to c Th e dyadic I c c c is a true quadr an t al versor for all vectors Th e powers of these dyadics behave like t h e powers of t h e imaginary unit V l l as may Applied to the b e seen from the geome t ric interpretation unit vectors i k Th e , x . . — , , j . , i x I = k j— I > < i Th e vector (a b) x x jk , e tc ( . 49) in skew multiplication is equivalen t I in direct multiplication 3 x b to . — a ( r a — a x b) ( x a x b) r -(b a — b) a b 0 5 ( ) -r 1 5 ( ) b) . complete dyadi c may b e reduced to a sum of three dyads of whi ch the anteceden t s among t hemselve s and t h e consequents among t hem s elves each form a right -handed rectangular system of three unit vectors and of whi ch the scalar c o e flic ie n ts are all positive or all negat ive A . 3 5 ( ) A n in c o m s is called the normal form of the dyadic Thi u t one or more of b dyad c may reduced to th s form i b e i l t e e p the c o e fli c ie n t s are z ero Th e reduc t ion is unique in case the constants a b c are difi e re n t In case t hey are not ferent the reduc t ion may b e accomplished in more than di f one wa y An y self-conjugate dyadic may b e reduced to the normal form . . , . , . in whi ch the consta nt s a , b, c y 55 ( ) are not necessaril positive . VEC TOR A NAL YSIS 3 28 i yd double do t and doub le cross mul t iplicat on of d is de fi ned by the equations Th e a a b b i d= z c c d= a a > < b c o o s ( 56 ) d, 5 ( 7) b x d c a . double dot and doub le cross multiplication of dyadics is o b tained b y e xpanding the product formally accordi ng to Th e t h e di stri b utive law into a sum of products of dyads doub le dot and doub le cross multiplication of dyadics is com mutative but not associative O n e -half the doub le cross product of a d ad c O b i t sel f is called the second of O If Th e , . , y i . y . O2 l xm . ( 61 ) O n e -third of the double dot product of the second of O a n d O is called the th ird of O and is equal to the product of the scalar triple product of the antecedents of O an the scalar triple product of the cons equent of O d . é O; O O8 : O= [ a b c ] m l [ n ] . second of the conjugate is the conjugate of the second Th e third of the conjugate is equal to the third of t h e original dyadic Th e second and third of the reciprocal are the reciprocals of the second and third of the second and third of a dyadic Th e second and third of a product are the roducts of the seconds and thi rds Th e . . p . . T ( o le 0 ( 92 16 ( r $3 2 so 2 ¢2 . r. a . VECTOR A NAL YSIS 330 Prove the statements made in Art 1 06 and t h e con verse of the statements T Q then Show tha t if 9 is complete and f O Q 7 G ive the proof by means of theor O and T are equal developed prior to A rt 1 09 T O t hat D e n i t io n : Two dyadics such that O T 8 is t o say t wo dyadics that are commutative — are said to b e Sh ow tha t if any num b er of dyadi cs are h o m o ge ho m o lo go u s neons to one another any other dyadi cs whi ch may b e ob t ained from them b y addit ion sub traction and direct mul tiplication are homologous to each other and to the given dyadi cs Show al so that the reciprocals of homologous dya di cs are homolo —1 1 O gous J ustify the statement that if O T or T which are equal be called the quotient of O by T then the rules governing addition sub t raction multiplication and divi s io n of h o m o lo go u s dyadics are identical with the r ul es governi ng these opera t ions in ordinar algeb ra — it being understood that incomplete dyadics are analogous to z ero and the idemfactor to unity Hence the algeb ra and higher analysis of h o m o lo go u s dya di cs is practicall identical with that of scalar quantities Show tha t ( I x c ) O = c > 9 < T and ( c x I) O = e x O 10 Show that whether or not a b c be coplanar 6 . . i . . ' y . . fi . . . : , . , , , . ‘ o . , , , , , y , y . , . o . o . , . , and are copl a nar use the ab ove relation to prove the law of sines for the triangle and to ob t ain the relation with sc a lar coeffi cients which e xists between three coplanar vectors This may b e done b y multipl ing the e uation b a u nit normal to the plane of a b and c 11 . If a, b, c y . , 12 . , q y . What is the condition which mus t sub sist between the i coe ffi cients in the e xpansion of a dyadic into nonion form f L INEAR VECTOR FUNCTIONS 331 the dyadic b e self-conjuga te ? What if the dyadi c be anti self-conjugate , Prove the statements made in Art 1 1 6 concerning the numb er of ways in which a dyadic may be reduced to its normal form 13 . . . necessary and s uffi cient condition that an anti self-conjugate dyadic O b e z ero is t hat t h e vector of the dyadic Shall b e z ero 14 . Th e y . 15 . Show that if O be an dyadic the product O O0 is self-conjugate Show how to make use of the relation O 0 to 16 demonstrate that the antecedents and consequents of a self conjugat e dyadic are the same ( Art . x . . 17 . Show that (172 i O2 O3 2 O and y that if the double dot product O O of a d adic by its elf vanishes the dyadi c vanishes Hence ob tain the condi t ion for a linear dyadic in t h e form O2 O2 O -OZ f 1 9 Sh o w t h a t 18 . Show , . . . . 20 . T)3 = O3 + (172 Sh o w t h a t T + O : T2 + T3 : that the scalar of a product of dyadics is cha nged by cyclic pe rmutation o f the dyadi cs That is 21 . Sho w . ( O m a fi ? a O w a r - ( r o s . un C H AP TER VI R OTA TIO NS AN D STRAIN S y the foregoing chapter the a n a ly tic a l theor of dyadics has b een dealt wi t h and b rought to a state of completeness which is nearly fi nal for practical purposes There are however a numb er of new questions which presen t themselves and some old ques t ions which present t hemselves under a new form when t h e dyadic is applied to physics or geome t ry Moreover it was for the sake of t h e applica tions of dyadics that the t heor of them was developed It is then t h e o b ject of the present chapter to supply an e x tended application of dyadics to the theory of rotations and strains and to develop as far as ma appear necessary the further analytical theory of dyadics That the dyadic O may b e used to denote a transformation of space has already b een men t ioned A kno wledge of the precise nature of thi s tran s formation ho wever was not needed at the time Con s ider r a s drawn from a fi x ed origin and r as drawn from the same origin et now IN . , , y . . y , , . . , , . . r ’ ’ L , = Or . This equation therefore may be regarded as defi ning a trans format ion o f the points P of Space Situated at the terminus of ’ r into t h e point P situated at the terminus of r Th e origin remains fi x ed P oints in the fi nite regions of space remain in the fi nite regions of Space A n y point upon a l ine ’ , . . . r b e c o me s a nt i o p r ’ = b + x a = Ob + :c a O- . VECTO R A NAL YSIS 3 34 is important to notice t hat the vector 8 denoting a plane area is not transformed into the same vector 3 as it woul d This is evident from the fact that in b e if it denoted a line t h e latter case O acts on 3 whereas in the former case O2 ac t s upon 3 To Show that volumes are magni fi ed in the rat o of O3 to unity choose any three vectors (1 e f whi ch dete rmi ne t h e volume of a parallelopiped [ d e f] Ex press O wi t h the vec tors which form the reciprocal system to d e f as consequents It ’ . i . , , . , O c , . ’ f . i dyadi c O changes d e f into a b c ( whi ch are d fferent from the a b 0 a b ove unless d e f are equal to l m Hence the volume [ d e f] is changed into the volume [a b Th e , , , , ’ , , , O3 [a b c , , [a b d e f O [ ] 3 ] 0 , ] [ d e f] Hence ’ . . y ratio of the vol um e [ a b c ] to [ d e f] is as O3 is to un i t B u t the vectors (1 e f were a n y three vectors which deter mine a parallelopiped ence all volumes are chang ed by the action of O in the same ratio and this ratio is as O3 is to 1 Th e , . H , . . Ro ta tio n s a bou t Fi xe d P o in t a y Ve rs o rs . i necessar and su ffi cient cond tion tha t a dyadi c represent a ro ta tio n about some axis is that it be reducible to the form Th e o re m 1 2 5] j : Th e where and i systems of unit vectors et i L ’ ’ k’ , j , k dd are two right-han e . xi x + y j+ i ’ ’ yj z k z k ’ . 1 ( ) rectan gular Hence if R O TA TI ONS A ND S TRA INS 3 35 j is reduci b le to the given form the vectors i k are changed into t h e vec t ors i k and any vector r is changed fr om its po s ition relat ive t o i k into the s ame posi k tion relative t o i Hence b y the transformation no change of Shape is e ffected Th e s t rain reduces to a rota t ion k Conversely suppose the w hi ch carries i k into i b ody su ffe s no change of Shape — that is suppose it su b jected Th e vectors i k mus t b e carried into ano t her t o a ro t a t ion right-handed rectangular system of unit vectors et these he i k Th e dyadic O may therefore be reduced to the form O j ’ , j ’ r , j j , ’ ’ , j , ’ . . ’ , , j ’ ’ , j , . , , . ’ , , ’ , . ’ , L ’ . ’ fi De A n i t i on : +k k . y d adi c which is reducible to the form ’ ' ’ jj i i k k and which consequently represents a rotation is called a ve r s o r . conjugate and reciprocal of a versor are equal and conversely if the conjugate and reciprocal of a dyadic are equal the dyadic reduces to a versor or a versor multiplied b the negative Sign Th e o re m : Th e , L et y . ¢ = = i l l i 1 k k + + jj C l r l O —l — r OC . Hence the fi rst part of the theorem is proved second part let O 3 1 b j c k, O O a a + b b + c c o OC = =L . To prove the VE C TOR A N AL YSIS 336 q Hence ( Art 1 08) the antecedents a b c and t h e cons e uents Hence ( page 8 7) they a b 0 must b e reciprocal systems must b e either a right -handed or a left-handed rectangu lar system of unit vec t ors Th e left-handed system ma be changed to a right-handed one b y p re fi xin g the negative Then Sign to each vector , . , , . , y . . _ O_ third or determinant of a versor is evidently equal t o u nity ; that of the versor with a negative Sign to minus one Hence the criterion for a versor may be stated in the form Th e . , O 0 O0 = L 2 ( ) inasmuch as the determinant of O is plus or minus one if O OG = I it is only necessar to state th at if Or y , O o O0 = I, O is There a ve rs o r . are two geometric interpretations of the transforma tion due to a d adic O such that y $ ¢3 = l ¢ l 0 O= i i transformation due to O s one of rotation combined w th ’ re ection in the ori gin Th e d adi c i i + + k k causes a rota t ion ab out a de fi nite ax is it is a versor Th e negative Sign then reverses the direction of ever vector in Space and replaces each fi gure by a fi gure symmetrical to it with respect ’ ’ to the origin B y reversing t h e d rections of i and the syste m i k still remains right-handed and rectangular but the dyadic takes t h e form Th e fl . y ’ y ’ , j i . ’ j , , : . ’ — O jj ’ ’ k k, VECTOR A NA L YSIS 338 — jk) 4 ( ) . — ii = k k I . ij + . k j— Hence O=ii + y jk q (I c os I x i z — i i) + . s in i I x q 5 ( ) . i more generall in place of the axis any axis denoted If by the unit vecto r 3 be taken as the axis of rotation and if as b efore the angle of rotation about tha t a x is b e denoted b q the d adic O whi ch accomplishes the rotation is y y O= a a + c os , 6 ( ) I q ( how that this dyadic actually does a ccomplish the rotation apply it to a vec t or r Th e dyad a a is an idemfactor for a ll vec t ors parallel to a ; b u t an annihi lator for vectors perpendicular to a Th e dyadic I — s e is an idemfac t or for all vectors in the plane perpendic ul ar to a ; b u t an annihilator for all vectors parallel to a Th e dyadic I x a is a quadrantal versor ( Art 1 1 3 ) for vecto rs perpendicular to a ; but an annihilator for vectors parallel to a If then To S . . . . . r b e p a ra ll e l t o a (D . r a a r o z r . Hence O leaves unchanged all vectors ( or components of vectors ) which are parallel to a If r is perpendic ul ar to a . Hence the vector r has been rotated in its plane through the angle q If r were any vector in space its component parallel to a s u fi e rs no change ; but its component perpendicular to a is rotated about a through an angle of q degrees Th e whole vector is therefore rotated about a t hrough that angle et a be given in terms of i k as . . L , a a — a 2 1 - 1 1 1 a 0 0 j . , 0 1 a 2 ° rj + a 0 1 a 1 3 k R O TA TI ONS A ND S TRA INS 3 39 jk + a a 3 I x ki + 1 = 0i i a M O { + a 2 a jl a {a a ga z { 0 + { a + { a { a a 2 3 s z 1 ( 3 1 ( l (1 (1 2 2 a 3 — n a kj 2 a i j+ 3 on a - 2 , i k, z s a k k, k, (1 1 1 a a a a d -c o s — 008 3 (1 — 1 (l — a (l 1 ( + q) -d cos a cos 9} i i 2) c os sin q } i j q) q) — a — a 1 Sin k j qj sin q } 2 ki kj cos q ) + c o s q } kk . ( 7) be written as in equat ion ( 4) the vecto r of and the scalar of O m a y be foun d If O O . O8 = 1 + 2 c os q . ax is of rotation i is seen to have the direction of O the negative of t h e vector of O This is true in general Th e direction of the ax is of rotation of any versor is the negative of the vector of O Th e proof of this statement An y versor O depends on the invariant propert y of O may be reduced to the form ( 4 ) by ta king the direction of i Th e x, . . . x . VECTOR A NAL YSIS 34 0 coincident with the direction of the a xis of rotation After this reduction has b een made the direction of the ax is is seen B u t O is not altered b y the t o b e the nega t ive of O reduction of O to any particular form — nor is the ax is of rotat ion altered b y such a reduction Hence the direction of O the dire c t h e a x is of rotation is always coincident with tion of the negative of the vecto of O Th e tangent of one-half the angle of version q is . x x . . r sin 1 + . q c os -OX x/ Ox ( 8) 1 + OS q tangent of one half the angle of version is therefore determined when the values of O and O , are known Th e vector O and the scalar OS which are invariants of O deter mine completely the versor O et be a vector drawn in t h e di rection of the a x is of rotation et the magni t ude of be equal to the tangent of one -half the angle q of version Th e - . , x , . L Q , . Q . L . . Q ; 1 +O Q o ta n z éq . y vector determines the versor O completel Q will b e call ed t h e ve c to r s e m i t a n ge n t of ve rs io n B y ( 6 ) a versor O was e xpressed in terms of a unit vector parallel to the ax is of rotation Th e . . - . . O= + a a c os q (I — a a ) + s in l q x a . Hence if G be the vector semi-tangent of version Q Q . + c os + q a si There > < Q V e -a 1 0 ( ) is a more compact e xpression for a versor O in terms of the vector semi tangent of version et 0 be any vector in Space Th e version represented by Q carri es - . L . . c — Q X c in t o c i Q x c . VE C TOR ANAL YSIS 342 and c c -( I M ul t ip ly b y c I x 0) — H ence th e -1 . i (c dyad c Q carr ies the vector 0 Q x 0 into the vector 0 x o no matte r what the value of 0 Hence the d adic O determines the version due to the vector semi-tangent of version carries the vector 0 — 0 0 into Th e dyadic I + I X c I + Q Q ( ) y . Q Q . . x . — y <c Q > c ( Hence the d adic I+ I x Q 1 + Q I I ( Q x Q) “ carries the vector 0 Q x 0 into the vector 0 if c be p e rp e n dic u l a r to Q as has been supposed C onse uentl the d ad c . 1 + Q Q q , y y i ° r di produces a rotation of all vecto s in the plane perpen cul ar If however it b e applied to a vector x Q parallel to Q to the resul t is not equal to x Q Q . , , . R OTA TI ONS AND i S TRA INS Q 34 3 ob viate th s di ffi culty the dyad Q whi ch is an annihilator for all vectors perpendicular to may b e added to the nu m e ra t o r Th e versor O may then be written To Q . , . 1 + Q a + l ( 1 0) Q ) QQ ( I x Q) Hence substitut QQI — . in g ( Th 1 — e is may be e panded x 1 + Q noni on form in Q (1 + a 2— b a bi i ] 0 k. 2 — +b (1 c 1 1 28 If a L et . 2 — a 2 + (2 a Q a 2 b — a 2— 2 2 a )ik b + 2 0 (1 1 ) ) kk “ c 2 y is a uni t vector 3 d adic of the form O 2 ( 1 2) I a a a biq u a dra n ta l versor Tha t is t h e dyadi c O turns t h e points of space ab out the axis a through two right angles This may be seen by setting q equal to 71 in the general e xp ession for a versor is . , . r O= a a + c os q (I — a s ) or it may be seen directly from geometrical cons iderations Th e dyadic O leaves a vector parallel to a unchanged but re verses ever vector perpendicular to a in direction Th e o re m : Th e product of two biquadrantal ve rs o rs is a ve sor the ax s of whi ch is perpend cular to the axes of the . y r . i i VECTOR A NAL YSIS 344 b iquadrantal and the an gle of which is twice the angle from the axis of the second to the ax is of the fi rst et a and b be the ax es of two bi uadran tal ve rs o rs Th e product ve rs o rs L q 9 y — I 2 b b ( ) : o . . s e — 2 ( I) y is certainl a versor ; for the product of an tw o ve rs o rs is a versor Co nsider the common perpen di c ul ar to a and b — 2 I reverses this perpendi cular a iqua rantal versor a d b Th e in direction ( 2 b b — I) again reverses it in direction and con se q uently b rings it b ack to its origin al position Hence the produc t 9 leaves th e common perpendicula r to a and b u n changed 9 is therefore a rotation about this line as axis . . . . . Qa = . Th e a . 2 b b ( — I) o 2 ( cosine of the an gle from -Q o e 2 b z -a b o I) — 3 3 a — a o a to a Q . o = 2 (b a o a o = 2 bb o a — a. is a ) 2 — 1 = cos Hence the angle of the versor Q is equal to twice the angle from a to b Th e o re m : Conversel any given versor may be ex pressed as the product of two biquadrantal ve rs o rs of which the ax es lie in the pla ne perpendicular to the ax is of the given versor and include between them an angle equal to one half the angle of the given versor Fo r let I? b e the given versor et a and b be unit vectors perpendi cular to the ax is — Q of thi s versor Furthermore let the angle from a to b b e equal to one ha lf the angle of this versor Then b y the foregoing theorem . y . , L . . . x . . O= ( 1 4) 2 ( bb resolution of ve rs o rs into the product of two b iq u a d rantal ve rs o rs a fi o rds an immediate and Simple method for compounding two fi nite rotations about a fi x ed point et O and T be two given ve rs o rs et b b e a unit vector per Th e . L . L VECTOR A NA L YSIS 34 6 O8 = 4 T= 4 c h o c ( T o O=4 c T ( o 4 z a o h) c b o 2 (o b) — c a O) x = 4 2 xb Hence a o QZ X Q I a b 2 — x c ) 1 z — a 1 + I, a a , . c ’ b -b ' ' B h b x Q2 ’ — c c e + I, x b, -a c 1, — 2 0c c o a 2 — 2 b b — b Tx = 4 Ts a o c c [ -c a a o b c ] b b b -c ° ut b Hence . a < Q1 Q2 > Hence Ql 0l , Q2 ( O b b Hence Hence o c 1 QI e -b b a o o c e h b o x b) a + > < Q, s a a -b a Q2 x a xb e a . b b . c a . c a o b b o c a . b b Q2 , QI Q" A T Q 2 1 11 ' ‘ . 1 b . . b c ROTA TI ONS AN D S TRA INS 34 7 This form ul a gives the composition of two fi nite rotations are b o t h in fi n it e s i If the rotations b e in fi nitesim a l Q 1 and mal N eglecti ng in fi n it e s im a ls of the second order the for mula reduces to . . Q3 QI i Q2 ‘ " . i in fi nitesimal rotations comb ine accord ng to the law of vector addi tion This demonstrates the paral lelogram law for angul ar velocities Th e subject was treated from di fferent standpoints in Arts 51 and 6 0 Th e . . . . Cyc li c s Righ t Te n s o rs , To n ic s , , If the dyadi c O a nd Cy c lo to n i c s y be a versor it ma be written in the form ( 4 ) — k q ( j s in + k j) . axis of rotation is i and the angle of rotation about th at ax is is q et T b e another versor with the same ax is and ’ an angle of rotati on e ual to q Th e L . T y : q ii + c oS q ’ . ji C + k k) + s in q ’ — k ( j jk ) . M u ltipl ing O T : T O : ii (jj+ c os k k) jk ) ( 1 6) > This is the res ul t which was to be e x pected the product of two ve rs o rs of whi ch the ax es are coincident is a versor with the same ax is and with an an gle equal to the sum of the angles of the two gi ven ve rs o rs If a versor b e multiplied by itself geometric and anal tic cons iderations alike make it evident that sin ( c ' k ( : +g — . y . , — and s in n — k q ( j jk) jk) , . VECTOR A NAL YSIS 34 8 O k j n jj the o t her hand let Ol equal — T k hen j O i i cos sin O q I ( and kk; q e ual O2 . ” q product of i i into either OI or O2 is z ero and into itself is Hence O rn O cos s O ii q 2) q I ( 1 l O O n cos O sin c os O ii q q q , l , Th e ii . ” " ” : : ” ” ” fl- " dyadic OI raised to any po wer reproduces itself OI OI Th e dyadic O2 raised to the second power gives the negative of OI ; raised t o the third power t h e negative of O2 ; raised to the fourth power OI ; raised to the fi ft h power O2 and so on ( Art Th e dyadic OI multiplied by O2 is e ual to Hence O2 Th e “ . . , q , , . . O =ii + ” B cos q O1 + " n c os " ‘ 1 q a si 2 ut E quating for O and OI O2 in these two e xpressions " c os n s rn n coe ffi cients of q = n q — c os n cos ” ‘ 1 q ( 2! n q s rn q y Thus l n — ( n — ) cos 1 — 2 the ordinar expansions for cos n q and Sin n q are ob tained in a manner ver Similar to the manner in which they are generally obtained Th e e xpression for a versor may be generali z ed as follows et a b c b e any three non -coplanar vectors and a b the reciprocal system Consider the dyadic y . L . , ’ , , ’ , . O : a a ’ + cos ’ q (b b + c c ’ ) + s in q (c b ’ 1 7 ( ) VECTOR A NAL YSIS 3 50 a displacement of t h e radius vector r an e llip tic ro ta tio n th r ough a sector q from its si m ilarity to an ordi nar rotation of which it is the projection De n i tio n A dya di c O of the form th e who le e llip s e Such 1 i s to 2 7 r q may b e called a s . fi y . O a a : ’ + cos ’ q (b b + c c sin ’ ) ( g c b ’ — b c ’ 1 ( 7) ) is called a c yc lic dya di c Th e versor is a Special case of a cyclic dyadic It is evident from geometric or analytic considerations that the po wers of a cyclic dyadic are formed as the powers of a versor were formed b y mul t ipl i ng the scalar q by the power to which the dyadic is to be raised . . y , , . O If the " a a cos ’ n q (b b ’ c c ’ ) sin n q ( c scalar q is an integral sub-m ul tiple of 2 2 b 7r, ’ , ' 71 9 i i th at is f p it s possible to raise the dyadic O to such an integral ower namely the power m that it becomes the idemfactor , , , O m =I may then be regarded as the m t h root of the idemfactor In like manner if q and 2 n are commensurab le it is possi b le to raise O t o such a power that it becomes equal t o t h e idem factor and even if q and 2 are incommensurab le a po wer of O may be found which difi e rs by as little as one pleases from t h e idemfac t or Hence any cyclic dyadic ma be regarded as a root of t h e idemfactor O . 71 y . g . i d e n t t h a t fix i n t h e re s u l t o ft h e a ppli c a t i o n o f T t o a ll ra dn ve c t o rs in a n e llip s e p ra c t i c a l l y fi x e s it fo r a ll ve c t o rs in t h e pl a n e o f b a n d c Fo r a n y ve c t o r in t h a t pla n e m a u l t ipl e o f a ra di us ve c t o r o f y b e re a rd e d a s a s c a l a r t h e e l lips e 1 It is ev . g m . fi De R O TA TI ONS A ND n it i o n : O= w here transform Th e a ii + b S TRAINS ation rep resented jj+ c the by ( 1 8) kk are posi t ive scal ars is called a p u re s tra in dyadic itself is called a right t e n s o r A right te n sor may b e facto red into three facto rs b, a , 3 51 c Th e . . b k k i i i i c + + j + j j j j j ) ( ( Th e o rder in which these fa ctors occur is immate ial transformation o z a . r (i i + jj+ c k k) o k k) . . Th e r is such that the i and jcomponents of a vecto r remain u n altered but t h e k-c omponent is altered in the rati o of c to 1 Th e transforma t ion may t herefore b e described a s a s t retc h or elongation along t h e dir ec t ion k If the cons t ant c is greater t han uni t the elongation is a tru e elonga t ion : b u t if c is less t han unit the elongati on is really a compression for the ra t io of elongation is less t han uni t y B etween these two cases comes the case in which t h e constant is uni t Th e lengths of the k-c omponents are then not a ltered Th e transformation due t o t h e dyadi c O may b e regarded the successive or simultan eous elongation of t h e com as n ts of r parallel to i and k respectively in the ratios n e o p a to 1 to 1 c to 1 If one or more of the cons t an ts a b c is less than unit t h e elongation in tha t or t hose di rections If one or more of the constants is b ecomes a compression uni t y components parallel to that di rection are not altered Th e di rections i k are called t h e p ri n c ip a l a xe s of the s tra in Their directions are n o t altered b y t h e stra in whereas if the cons t an t s a b c be different ever other direction is al t ered Th e s c alars a b c are known as t h e p ri n c ip a l r a tio s of . ' y y . , y . . . , b , y , j , , . , . , , , tio n , , . , , e l o n ga j . y , , . it wa s seen that any complete dya di c reducib le to the normal form In Art . . 115 wa s SS VECTOR A NAL Y I 3 52 y where a b c are positive constants Thi s expression ma be factored in to the product of two dyadi cs , . , . o ( :t : a ’ ’ ' i i + b jj+ ’ a 1 9 ( ) ) + k k) factor Th e k k ’ ' i : c ' i i ’ + ’ ' k k jj which is the same in ei t her metho d of factoring is a versor k T i turns the vectors i k into the vectors his j It versor may b e represented b y i ts vector semi-tangent of , j . ' ' , , , ' . ve rs io n a s 1 1 + JJ+ ’ k k 1 +i o 1 ’ ’ + JJ+ k o k other factor Th e a ii + bj j+ c kk is a right tensor and represents a pure strain In the fi rst ' case the strain has t h e lines i j k for principal ax es : in the second i k In b oth cases the ratios of elongat ion are the same — a to 1 to 1 c to 1 If the negative sign occ u rs b efore the product the version and pure strain mus t have associated with them a reversal of direc t ions of all vectors in space — t hat is a perversion Hence The o re m A n y dyadic is reduci b le to the product of a versor and a right tensor taken in either order and a pos itive or negative sign Hence the most general transformat ion representab le b y a dyadic consists of the product of a rota tion or version ab out a de fi nite ax is through a de fi nite angle accompanied b y a pure strain either with or without pe rver sion Th e rotation and strain may be performed in either order In the two cases the rotation and the rat ios of elonga t ion of t h e s t rain are the same ; b u t the principal a x es of t h e strain differ according as it is performed before or after the . , , j , . , , , , . b . , . . . ’ ’ , 3 54 A NA L YSIS VE C TOR b componen ts are stretched in the rati os a to 1 to 1 c t o 1 If one or more of t h e cons t ants a b c are negative the com b c are re n e n t s parallel to the correspon di ng vector a o p versed in direction as well as changed in magnitude Th e t onic may b e fac t ored into thr ee factors of which each s t re t che s t h e components parallel to one of the vectors a b c b u t leaves unchanged the components p a rallel to the other , , . ' , , , , . , t wo , . value of a ton ic O is not altered if in place of a b c any three vec t ors respectively collinear wi th the m b e s u b sti t uted provided of course that the corresponding changes which are nece s sar b e made in t h e reciprocal system a b c B u t with the e x ception of this change a dyadic which is e xpressib le in the form of a tonic is so e xpressib le in only one wa y if the constants a b c a re different If t wo of the constants say 6 and c are equal any two vectors coplanar with t h e corresponding vectors b and 0 may be substituted in place of b and c If all the consta nts are equal the tonic reduces to a constant multiple of the idemfa ctor An y three non -00planar vectors may b e taken for a b c Th e product of two tonics of which the a x es a b c are the same is commuta t ive and is a tonic wi t h these ax es and with scalar coeffi cients equal respectively to the produc t s of t h e correspondi ng coe f fi c ients of the t wo dya di cs Th e , , ~ y , ’ ' , , , , . , , . . , . , , , . O Th e o 97 : 7 11 -O = a 1 a 2 a a ’ 4 b b generaliz ation of t h e cyclic dyadic ( c b ’ — b c ’ ) ( 23 ) ’ . ROTA TION S AND S TRAI S 355 N non-coplanar where a b c are three vectors of which a b c is the reciprocal sys t em and where t h e quantities a b c are positiv e or negative scalars This dyadic may be changed into a more conve n ient form b y de t erm ining the posit ive scalar p and the positive or negative scalar q ( which may always b e chosen be tween the limits 1 7r) so that , , ’ ’ , , , , ’ , . b =p c os =p s in c q q . a nd Then O a a a : Th ’ + 19 cos b b q ( ’ c c ’ ) +p s in q ( o h ’ ( 25) is may be fact red into the product of three dya di cs o O= { a a (a a a ’ cos ’ bb ’ q (b b c c ’ ’ ) ( c c ’ ) a a ’ ’ +p bb +p sin q ( c b c c ’ ) ’ order of these factors is immaterial Th e fi rst is a tonic which leaves unchanged vectors parallel to b and 0 b u t stretches those parallel to a in the ratio of a to 1 If a is negati ve the stretching must be accompan ed b y reversal It leaves Th e second factor is also a to ni c in direction unchanged vectors parallel to a but stretches all vectors in the plane of b and c in the ratio p to 1 Th e th rd is a cyclic factor Vectors parallel to a remain un changed ; but radii vectors in the ellipse of whi ch b and c are conjugate se m i -di ameters are rotated through a vector such that the area of the vector is to the area of the whole ellipse as q to O ther vectors in the plane of b and 0 may b e egarded 2 7r adi vectors of the elli ps e a s scalar m ul tiples of the Th e . i . . . . i . . . r i r . 356 Defin i tio n O ’ e a a i A : + p SS VECTOR A NAL Y I y d adic whi ch is reducib le to the form cos b b q ( ’ c c ’ ) sin +p (c b q ’ ( 2 5) ow ng to the fact that it combines the properties of th e c clic d adic and the tonic is called a c yc lo to n ic Th e product of two c y c l o t o n ic s whi ch have the same three vec t o rs a b c as antecedents and the reciprocal s stem for con sequents is a third c cloto mic and is com b a c mu tative y y . , ’ , ’ ’ y y , , , . = a = a ’ O : 1 2 a a a a ill ' ’ +p1 c os +p2 cos -O = a + 19 1 q1 q2 a a a + 101 1 )s Re du c ti o n (b b ’ ’ g1 sin l +p1 pz c os + 92 ) (91 s in a di a D cs t o C f y o ( q1 (e h q2 ( c (c b ’ b ’ — — b c c ) ’ ) + q2) ( 26) ' n o n ic a b ’ l Fo rm s y general an dyadic O may be reduced ei ther to a toni c or to a c y c l o t o n ic Th e dyadi cs for whi ch the reducti on is impossible may b e regarded as li miting cases which may be represented to an desired degree of a pp ro xi mation by tonics or c y c l o t o n ic s From th s theorem the importance of the tonic and c cl o to nic whi ch have been treat ed as nat ural generaliz ations of the right tensor and the cyclic dyadic ma be seen Th e proof of the theorem including a d scus sion of all t h e special cases that may arise is long and somewhat tedious Th e method of proving the theorem in general however is patent If three di rections a b c may be found wh ch a re left unchanged by the application of O th en O must b e a to nic If only one such directi on can be found the e e x ist s a plane in which the vectors suffer a change such as that due to the c y c l o to n ic and the d adi c ndeed proves to be s uch Th e o re m : In . y i y . i , y . , . . , i , . , y i r . 3 58 An y A NAL YSIS VE C TOR value of which satis fi es this equation wi l x that y That (O x l )3 l be such 0 . is to sa the dyadic O x 1 is planar A vector per O n dic u l a r to its consequents is reduced to z ero ence e p leaves such a di rection unchanged Th e further di s c u ssion of the reduction of a d adic to the form of a tonic or a cyclo tonic depends merely upon whether the cubic equation in se has one or three real roots Th e o re m If the cu b ic equation , . . y H . . x $3 $8 4 x $2 3 2 “ 1 has three r e al roots the dyadic to a tonic : 1 O may in general be reduced . For let a ac : be the three roots of the equation O— a I, O — L = Th e . b I, O c y i d ad cs — c I are in general planar et a b c be respectively three vectors drawn perpendicular to the planes of the consequents of these dyadics . , , . (O I) a O, ( (D — b I) -b = 0, ( O a O a O b O -c a a b b, c & , the roots a b c are distinct the vectors 00 planar Fo r suppose m a n b If , , . (O — c I) a, are non R O TA TI ONS m O a — O m m ( . m a n a a, N O o O o b b = 0 — n c b = bb . . a (a m = Oor Cons e + c a S TRAI S AN D n (b n = Oor b = a quently if the vectors c . are coplanar the roots are not distinct ; and t herefore if the roo ts are distinc t the vectors a b c are necessarily non-coplanar In case the roots are n o t di s t inct it is still always possi b le t o choose three non coplanar vectors a b c in such a manner t hat the equa tions ( 30) hold This b eing so there e x is ts a system a b c reciprocal t o a b c and the d adic which carries a b c into a a b b c c is the to n ic b, a, c , , , . , - , , y , , ’ , , . , ’ , ’ , , , O= The o re m If the : a a a ’ bbb ’ c c c . q cub ic e uation x2 y OS + cc (Dz s y has one real root the d adic O ma in general be reduced to a cycloto mic Th e cub ic equa t ion has one real root This must b e posi tive or negative a c cording as O3 is positive or negative et the root be a D ete rmine a perpendicular to the plane of the conse uents of O a I . L . . q . . ( D ete rmine at! O — a ( O I) o a = 0 . also so tha t a ’ — a I) = O and let the lengths of a and a be so adjusted that a a l Th s cannot be accomplished in the special case in whi ch a i ’ ' z . S L et VECTOR A N AL Y IS 3 60 and a are mutual ly perpendi cular the plan e perpendicular t o a ' . b y be an vector in ’ . 8” O— ( I) a b = 0 0 . Hence ( Hence O h is I) b is perpendic ul ar to a 2 3 m O perpendicular to a In a si ilar manner b O h and 3 1 O b etc will all be perpendicular to a and lie in O b one plane Th e vectors O h and b cannot be parallel or O wo ul d have the direction b as well as a unchanged and thus t h e cubic woul d have more than one real root z Th e dyadic O changes a O h b into O a O b O b re Th e volume of the parallelopiped s p e c t ive ly O— a ’ o o . ’ . — o , , ’ " . , , , o . . o , , o , , . O [ B c OZ a O O h o Oa ut Hence a a a b) a o b] h . . = O3 a vectors O h O h b all lie in the same plane Their vec t or products are parallel to a and to each other Hence Th e Z o o , . , ’ a Inasmuch Z O ( as and a x( h) o b 1 =p 2 O l ‘ vectors b 2 2 a -1 bz l z b3 + b1 = 2 b 1 + b _1 2 n n b2 b b x b . ( 3 2) . p z 2 and b 1 are parallel b O z ‘ O ‘ Oz 2 , n b1 L et . . b1 + b2 = 2 b -1 + b _2 — 2 o -b bl x bz, b2 + b = 2 Then O3 b _2 _ p b bs Th e = O3 Ob 1 o b) , p b3 = p O have the same sign let OS “ . n n b3 b b, etc etc . . 3 3 ( ) SS VE C TOR A NAL Y I 3 62 There i remain two cases in wh ch the reduction s impossible as can b e seen b y look ng over the proof In the fi rst place if the c o n s t a n t n used in the reduction to cyclo tonic form b e i 1 the reduction falls through In the second lace if the plane of the antecedents of i 1 i , . p . O— q I a r l and the plane of the conse uents a e perpendicu ar the ’ vectors a and a used in the reduction to c y c l o t o n ic form are ’ perpendicular and it is impossible to determine a such that a a sha l l be unity Th e reduction fall s through ’ . . If = i ” L et b _l + b 1 = i 1, bl b _1 Choose 0 y Consider the d a di c b1 z T= T ' ill o 2 b =b — — O = O a a a . . (b b p ’ b -b , = O h p c c c ’ , O ~ a a a b _l +p — Hence o b = p b i—p h 1 ’ o a a o z h “ 2b ’ ) +p -c . c bl ( 3 7) . Th e y transformation due to this dyadic may be seen best b factoring it into three factors which are independent of the order or arrangement O ( : a a a ’ ’ + b b + c c ’ ) ( 1 In In t h e s e one case dya d i t he s e wh i c h t wo mg a nd i a t wo t wo cs o c as e s o f th e ro o o ft h e ro o t s o f be se e n a re ts a e d re dis t i n c t Th e in a ry p a rt o f t h e . bb ’ mg d q lil mi i g lyi g b i m gi y i i li m i g d k i g pl i m gi y i pp i g il it w l ma y be ft h e a a ’ a r e re t as a na r m a y be t wo a c a n a nd na r h bi c e qu a t io n o th e r c a s e c a se s t h e to n de re ro a c cu in t h e a nd ua t th e t o n tha t th e ar as ta ro o ts o ft h e n C n c in wh n c e h a s t h re e t h re e t we e n t h e c h a ce y c l o t o n ic e a c h o the r . of a ll t h e m re a t he c W) l . ro o t s . Th us y c l o t o n ic in ro o mg t s a re re a l i th e r by t h e p u re be c o i n z e ro o r by e S TRAINS R O TA TI ONS A ND 3 63 Th e fi rst facto r represents an e l ongation in the dir ection a in a ratio a to 1 Th e plane of b and c is undisturbed Th e second factor represents a stretching of the plane of b and c in the rat o p to 1 Th e last factor takes the form . i . . I+ 1 ( + 1 ( + b c (I o ) ’ ) c y h ’ ma , ma cc b M c b = ’ ) zc b a o ' a s, cc c . d adic of the form I o h leaves vectors parallel to a and c unaltered A vector x b parallel to b is increased by the vec tor c mul tiplied b y the ratio of the vector a xb to b In other words the trans forma t ion of points in Space is such that the plane of a and c remains fi x ed point for point b u t the points in planes parall el to that plane are shift ed in the di rec t ion c by an amount proportional to the di stance of the plane in which they lie from the plane of a and 0 A dyad c reduci b le to the form Defin itio n A ’ . . i . I+ is called a c dyadi c or transformation which it causes general dyadic is l and the geometri ca calle a s h e a r Th e more s h e a re r s h e a ri n g ll h ’ ll d . ( 3 7) i wi al so be ca ed a shearing dyad c or s he a rer Th e trans formation to which it gives rise is a shear combined with elongations in the direction of a and is in the plane of b and c — 1 instead of n If n + 1 the result is much the same Th e dyadic then b ecomes . " . . , O O: (a a a ’ bb : ’ a a a c c ’ — e h’ ’ ) {a a ’ - p (b b ’ (I SS VECTOR A N AL Y I 364 facto rs are the same ex cept the second whi ch now re p re sents a stretching of the plane of b and 0 comb ined with a reversal of all the vectors in that plane Th e shearing dyadic O then represen t s an elongation in th e direction a an elonga tion comb ined with a reversal of direction in the plane of b and c and a shear Suppose that the plane of the antecedents and the plane of the consequents of the dyadi c O — a I are perpendi c ul ar et these planes b e taken respectively as the plane of and k and the plane of i and k Th e dyadic then takes t h e form Th e . , . , j j O coeffi cient B must vanish — a I — B I= L . O Th e . ( Fo r . B i + A — y other wis e the d adic — j E h) q is planar and the scalar a B is a root of the cubic e uation With this root the reduction to the form of a tonic may b e carried on as b efore N othing new arises B u t if B vanishes a new case occurs et . . This L . may be reduced as follows to the form a where a -b ’ z a Square T -c ’ Tz z b z — ’ b + b c ’ o o c z ’ A D ki = a nd a c b ’ . H nce must be chosen parallel to k ; and Th e dyadic T ma then be trans formed into e . a y T=AD k — 0i + Dj A D A D k, b =A j c + A ji ’ , . l parall e to i . VECTOR A NAL YSIS 366 i i more systematic treatment of the var ous k nds of dyadics which may arise may be given by means of the Hami lton -Cayle equat ion A y O Os I = 0 Ts O + O 2 5 O 2 3 and the cubic equation in x2 Os 3 — x x + Oz g x — i O3 = 0 . m are the roots of th s cubic the Ha ilton Cayle b equation may be written as If a , , c ( O — a -( O I) y 4 0 ( ) yy If, however, the cubic has only one root the Hamilton -Ca le equat ion takes the form (O — a 2 I) ( O 4 1 ( ) general the Hamilton Cayley equation which is an equ a tion of t h e third degree in O is the equation of lowest degree which is satis fi ed by O In general therefore one of the above equations and the corresponding reductions to the tonic or c y c l o t o n ic form hold In special cases however the dyadic O may satisfy an equation of lower degree That equation of lowest degree which may b e sat is fi ed b y a dyadi c is called its c h a ra c te ris tic equation Th e follo wing pos sibilit ies occur In - . . , , . . — O ( O ( — a I) O ( a I) . o ( 2 — O O ( — O ( — a a b I) 2p I) O (O o c os I) O ( (O ( — — -( O O ( q b I) 2 R OTA TI ONS AND S 3 67 TRA INS b e the rst c se the dyadic is a tonic and may reduced a fi In to the form O= In the second case reduced to the form ’ a a a + bb b ’ + dyadi c is a the c c c ’ . l o to n i c y c c and may In the third case t h e dya di c is a simple shearer and may reduced to the form O= a a a ’ + b( b b ’ + c c ’ ) + c be be ’ b . the fourth case the dyadic is again a tonic Two of the ratios of elongation are the same Th e following reduction may b e accomplished in an in fi nite numbe r of ways In . . . O= a a s ’ b( + bb ’ + the fi fth case the dya di c is a complex shearer and may be so e xpressed that In T : the six th case the dyadic may be reduced to the form In O= a I + c b ' — . a ( is a s ' again a simple shearer which + bb ' + c c ' ) + c b ’ . the seventh case t h e dyadic is again a tonic which ma reduced in a doub ly infi nite numb er of ways to the form In These y be seven are the only essentially di fferent form s which a dyadic may take There are then only seven really different kinds of dyadi cs three tonics in which the ratios of elonga tion are all difi e re n t two alike or all equal and the cyclo tonic toge t her wi t h thr ee mi t ing cas es the two simple and the one complex shearer . li , . , , , VECTOR 3 68 Su mm a ry r ANA L YSIS f o Ch a p te r VI t ansformation due to a dyadic is a linear h o m o ge n e o u strain Th e dyadic itself gives the transformation of the points in space Th e second of the dyadic gives t h e trans form ation of plane areas Th e third of the d ad c gives the ratio in which volumes are changed Th e . y i . . . D l = ¢ 3 1k y necessary and suffi cient condition that a d adic re p re sent a rotation about a de fi nite ax is is that it be reducible to the form Th e or tha t or th a t OO0 = I O3 = + 1 T T0 = I T3 > 0 0 y y necessar and suffi cient condi tion that a d adic repr o sent a rotation combined with a transformation of reflection b which each fi gure is replaced by one symme t rical to it is that Th e y O= or ’ + k k) OOC = th a t O3 . or that O O0 1 — O3 0 r 3 ( ) . dyadic of the form ( 1 ) is called a ve sor one of the form a perversor If the a x is of rotation of a ve sor be chosen or the i-a xis the versor reduces to A r . O = ii + cos j+ k k ) q (j O = ii + If an y unit vector O= Th e a a 9. q ( 4) - k k q ( j j) (I 5 ( ) is directed along the ax is of rotation + r cos sin c os q (I i axis of the ve sor coinc des in i on wi h di rect t Ox . 37 0 VECTOR A NAL YSIS y duct of two c clic dyadics whi ch have the same antecedents ’ b c and consequen ts a b c is obt ained b y adding their a angles q A cyclic dyadi c may b e regarded as a root of the idemfactor A dyadic reducib le to the form ’ , ’ , . . l O= ii + bj j+ a c kk where a b c are positive scalars is called a right tensor It represent s a stretchin g along the principal a xis i k in the ra t io a to 1 to 1 c to 1 which are called the principal rati os of elonga t ion Thi s transformat ion is a pure strain A n y dyadic may b e e x pressed as the product of a versor a right tensor and a positive o r n e ga t ive sign , , , b , j . , , . . , , O i : (a . . ’ i i ’ ’ ’ b jj c ’ k k ’ ) (i ’ i + ’ jj k ’ k) or ( 1 9) Consequently any linear homogeneous strain may be regarded as a comb ination of a rota tion and a pure strain accompanied or unaccompanied b y a perversion Th e immediate generali z atio n s of the right te nsor and the cyclic dyadic is to the tonic . O a a a b ’ and c y c l o t o n ic bb ’ c c c ’ — b c or ( 23 ) ) — where An y = + P Vb + z c z and ta n y l q 2 — p — b c ’ ) ( 2 5) b b + p d adic in general may be reduced either to the form and is therefore a tonic or to the form and is t herefore a c c l o t o n ic Th e con di tion that a d adic be a y tonic is that the cubic equati on y , . x 3 «9s + a ,, x a8 = o R OTA TI ONS AN D S TRA INS 371 shall have three re a l roots Special cases in whi ch the eduction may b e accompl ished in more ways than one arise when the equation has equal roots Th e condi t ion that a dyadi c b e a c y c l o t o n ic is that this cubic equation shall ha ve only one real root There occur two li mitin g cases in which the dyadi c cannot b e reduced to c y c l o to ni c form In thes e cases it ma be written as r . . . y . O = a a a ’ + + p c b ’ and is a simple shearer or it takes the form , 3 8 ( ) y l and is a comple x shearer Dyadics ma be c a ssi fi ed accord in g to their characte ri stic equa t ions . O ( (O — a I) — a l) b I) ( o O — c I) = O (O y O ( I) a O— a ( I) O ( b O -( O — a 0 simple shearer bl) = 0 special tonic I) 2 I) 3 0 complex shearer 2 0 special simple shearer (O a I) O a I) ( tonic c cloto mic 2 g ( -( O — 0 pecial tonic S . C H A P TER VII M ISCELL AN E OU S A PPLI CA TIO NS If O Qu a dric Su rfa c e s y q i be an constant dya di c the e uat on r O r const . y is quadratic in r Th e constant in case it be not z ero ma be divided into the dyadic O and hence the e uation takes the form 1 O r r r O 2 r = 0 ( ) q , . , , . y dyadi c O ma be assumed to be self-conj u gate Fo r if T is an antiself-conjugate dyadic the prod u ct r T r is identicall y z ero for all values of r Th e proof of this state ment is left as an e x ercise B y Art 1 1 6 any self-conjugate dyadi c is reducible to the form Th e . , . . . O_ i i i a 2 i b a m Hence the equation r O k k jj , i . c y r a 1 y represents a quadric surface real or imaginar Th e dif ferent cases which arise are four in number If the signs are all positive the quadric is a real elli psoid If one sign is negative it is an hyperboloid of one shee t ; if two are . . , . VECTOR A NAL YSIS 3 74 Hence the most general quadratic e xpression may be reduced to O r where O is a constant dyadic A a constant vector a nd C a constant scalar Th e d adic may b e regarded as sel f conjugate if desired A make a change of or g n To be rid of the linear term ’ b replacing r b y r t ’ y . , , r . y ii , . ( ’ — r ’ r O o equal . ’ ( — r o + l Since O O t) t ’ r ’ r — t + ) O o o ( ’ — r ’ — r r ’ t) -O o o A+ -O t + t o t A— t o r is se f-conjugate the second and thi rd terms a e Hence ' r If now O . -A O is complete the vec t or t may b e chosen so that 1 § A= O o t or t Q Hence the quadric is reducible to r ’ O l — t he O r -A . central form const ’ 1 . case O is incomplete it is u n iplanar or u n ilinear O is self -conjugate If A lies in the plane of O or in of O as the case may be the equation In . 1 2 A= w o t is solub le for t and the reduction to central form is still pos sib le B u t unless A is so situated the reduction is impossib le Th e quadric surface is not a central surface Th e discussion and classi fi cation of the var ous non-central quadri cs is an interesting e x ercise It will not be taken up he re Th e present ob ject is to develop so much of the theory . . . . . i Q UADR I C S URFA CE S 3 75 of quadr ic surfac es as will be useful in applications to mathe m a ti c a l physics wi t h especial reference t o non -iso t ropic media Hereaft er therefore the central u a dr cs and in parti cul ar the ellipsoid will be discus sed Th e tangent plane may be found b di f ferentiation q i . y . r dr Since O O o + r o -r = O o r O o 1 . . dr = 0 o . is self conjugate these two terms are e qual and - O dr r O . increment d r is perpen di cul a r to O r Hence O r is normal to the surfac e at the e x tremity of the vector r et this norm a l be denoted b y N and let the u nit normal be u Th e L . . . N = T L et r ~ be the vector dr awn fr om the origin perpendicular to the tangent plane p is parallel to n Th e perpendicular distan ce from the origin to the tangent plane is t h e square root of p p It is also equal to the squ are root of r p p . . . Hence r o = p r P P B ut H ence inasmuch as r c os r p ' ‘ P o = p p p . P _ P P -O -r o r ‘ P -N z l . y q and N are parall el the are e ual , O -r = N . VECTOR A NAL YSIS 37 6 page 1 08 it was seen t hat the vecto r which ha s the dire c tion of t h e normal to a plane and which is in magnitude equal to t h e reciprocal of the di stance from the origin to the plan e may be ta ken as the vector co ordi nate of that plane Henc e the above equation shows th at O r is not merely normal to the t angent plane b u t is also the co ordi nate of the p lane That is the length of O r is the reciprocal of the distance from the origin to the plane tan gent to the elli psoid at the e xtremity of the vector r Th e equation of the ellipsoid in plane c o tirdin a t e s ma be found b eliminating r from the two e uations On . ' , . , y . y q r O o O r Hence r = -O -e l “ -O l, z =N r o . -O -O -O I -H N o I ‘ -r . -l N z ‘ . O 1 . q Hence the desired e uati on is N O o ‘ 5 — Let a 2 I -N = 1 . ’ + b ii + s j+ 0 2 kk . and where a v w are the reciprocals of the intercept s of the plane N upo n the ax es i k Then the ellipsoid may be written in either of the two forms famili ar in Cartesian geometry , , , . j , . 3 78 r vecto s dyadi c O Th e th e VECTOR ANALYSIS Let . a r vecto s Th e b, a, =O ’ ’ a , b Fo r ’ , a a o ’ 0 o are changed into c b = O ' , O o c O O b, O o a c y b ’ c a, b, c . = 1, a -b dyadic h, a, form the system reciprocal to c Th e o O -O -c = 1 = 0, ’ c o b -0 ’ O a o , =b o O c o = 0, = O . may be therefore e x pressed in the forms O O ’ b b a a + b b a s —l — ’ ’ ’ ’ c c ’ c c . for convenience the three directions a b c be called a system of three conjugate radii vectors and if in a si m ilar manner the three tangent planes at their e x tremities be called a system of t hree conjugate ta ngent planes a number of geometric theorems may b e ob tained from interpreti ng the invariants of O A system of three conjugate radii vecto s may be obtained in a doubly infi nite number of wa s Th e volume of a parallelopiped of which three concurrent edges con sti t ute a system of three conjugate radi vectors is constant and equal in magnitude to the rectangular p arallelo piped constru cte d upon the thr ee semi -ax es of the ellipsoid Fo r let a b c be any system of three conjugate ax es If , , , , , y i . r . . , Th e . , determ inant pendent of th e or third of O is an in variant form in which O is e xpressed ‘ I . Q UA DR I C S URFA CES B ut 3 79 if O3 Hence [ “ a 1 a b c 2 ] This b 2 a b c ” . c. demons trate s the theorem In l ike manner by inter 1 O O and b it is possi le to show that O re t in S S p g 3 Th e sum of the s uares of the radii vectors drawn to an ellipsoid in a system of three conjugate di rections is constant and equal to the sum of the Squares of the semi-ax es Th e volume of the parallelopipe d whose three concurrent edges are in the di rections of the perpe n di culars u pon a system of thr ee conjugate tangent planes and in magnitude equal t o the reciprocals of the d stances of those planes from the center of the ellipsoid is constant and equal to the reciprocal of the parallelopiped co n structe upon the semi -ax es of the ellipsoid Th e sum of the squares of the reciprocals of the thre e per n di c u l a rs dropped from the origin upon a system of three e p conjugate tangent planes is constant and equal to the sum of the squares of the reciprocals of the semi -ax es If i k b e three mutuall perpen di cular unit vectors . q “ , , . , i d , . , L et j y , O aa , ‘ l o i . + j-O " 1 j+ be th ree radii vectors respectivel parallel to i k a, b, y c , a -O Hence i a a o o j , in k o O ' 1 o k the ellipsoid drawn . =b O -O o i + o a j o b O -j -O -b k + c . O -O -k o c the three terms in this e xpression are the squares of the reciprocals of the radii vectors drawn respectively in the i ence : k d rections B ut i . H , j , VECTOR A NAL YSIS 3 80 l y sum of the squares of the reciproca s of three mutuall perpendicul ar radii vectors in an ellipsoid is constant An d in a similar manner : the sum of the squares of the p e rp e n dic u l a rs dropped from the origin upon three mutuall p e rp e n di c u l a r t angent planes is constant Th e equation of the polar plane of the po nt deter 1 a mined b y the vector is O3 s a = 1 1 ( ) Th e . y i . . let s be the vector of a point in the polar plane Th e vector of any point upon the line which joins the termin u s of s and the termin u s of a is Fo r . y s a + :c a + g / is point lies upon the surface If th s y ar + ma t y 2 s - w a ' w z l If the terminus of s lies in the polar plane of a the two values of the rati o m y determined b y this equation must be equal in magn it ude and opposite in Sign Hence the term in m y vanishes Hence O a = 1 s . . o o is the desired equation of the polar plane of the terminus of a et a be replaced by z a Th e polar plane becomes L . . s s 1 re a mg It is l or i -O o a , z mm , l z 1 id e n t ly i a t e ri a l wh e t h e r t h e ina ry e llip so id o r h ype rbo lo id ev a -O -z a . c e n t ra m l qu a dri c d e t e r i n e d by 4»b e VECTOR A NAL YSIS 3 82 i is any point outside of the qu a dric and f all the tangent pla nes which pass t h rough A are drawn these pla nes envelop a cone This cone touche s the quadric along a plane curve the plane of the curve bein g the polar plane of the poin t A Th e equation Fo r let a be the vector dr awn t o the point A of any tangent plane to the quadric is If A , . . . 8 If this 0 0 1 r . plane contains A its equation is satis fi e d b y a Hence the condi tion s which must be satisfi ed by r if its tangent plane passes through A are . , a. r T o O l 1, ‘ 1 r . points r therefore li e in a plane r ( O a ) 1 whi ch on comparison with ( 1 3 ) is seen to be the polar p ane of A Th e quad ric which passes through the curve of intersection of thi s polar plane with th e given quadric and whi ch touches the quadric along that curve is Th e l ( If this T e o r passes through the point A ( Hence r ( O a O r o s O -r O -a . , — 1 r — 1 a O a ) ( ) — ( a o O a r transforming the origin to the point A this is easily seen to be a cone whose verte x is at that point et O be an self-conjugate dyad c It is e xpres sible in the form By L y . O= A ii + B B ’ , C jj+ i . Ckk are positive or negative scalars A < B O— B I= ( 0— B ) < . 0 kk — — A ii B ( ) . Further Q UADR IC S URFA CE S 3 83 ’ Then O B I=c + a )(o — a a c + c and =p a O= EI+ — a c = q . 1 1 4 ( ) 2 dyadi c O has been ex press ed as the sum of a constant multiple of the idemfactor and one ha l f the s u m Th e p q + q p ~ reducti on has assumed taci t ly tha t the cons tants A B C are different from each other and from z ero This e x pression for O is closely rela ted to the circular sections of the quadric sur fa ce O r r = 1 Th e , , . . Substituting the value of B L et O, -r -i- r -p r r r o r + q p be any plane perpendi cular to B O r 1 r = l becomes . n p . q n o r o By substitution r — 1 = O . This is a sphere b ecause t h e te rms of t h e second order a ll have th e same coe ffi cient B If the equation of thi s sphere be sub tracted from that of the given quadri c th e resul ting e q uation is that of a quadric which passes through the inte r section o f the Sphere and t h e given q u adric Th e difference . , . 18 q o r ( r o p Hence the Sphere and t h e qu a dric intersect in two plane c u rves l i ng in the planes = r r and 0 n p q y . o z . VECTOR A NAL YSIS 384 y i Inasmuch as these curves lie upon a sphere the are circles Hence planes perpendi cula r to p cut the quadr c in circles In like manner it may b e shown that planes perpendicular to Th e proof ma be conducted as q cut the quadric in circles follows : = 1 B r r + r r p q . . y . o . r is a radi us vector in the plane pas s ed through the center of the quadric perpendi cular to p or q t h e term r p q r van ishes Hence the vector r in thi s plane sat s fi es the equati on If i , . B r r o =1 r and is of constant length Th e section is therefore a ci cular section Th e rad us of the section is equa l in length to the mean semi—axis of the quadric Fo r convenience let the quadric be an elli psoid Th e con l s ta n t s A B T are then ositive h reciprocal d i c O C e a p ma be reduced in a simila r manner i . . . . y , . , . k Then dy O -l and d I = ff— d d — d ) + f ( ‘ 3 86 VECTOR A NA L YSIS i the rotat on the section is a circle In li ke manner consider the projection or shadow of the ellipsoid cast upon a plane parallel to the mean ax is b y a point at an infi ni te dis tance from t hat plan e and in a direction perpendi cular to it A s the elli p s oid is rotated ab out it s mean ax is from t h e position in whi c h the major a xis is perpendi cular to the plane of p ro j ec tion to t h e position in which the mi nor ax is is perp endi cul ar to t ha t plane the Sh a d ow and the projecting cylinder have t h e mean ax is of the elli psoid as one axis Th e other axis cha nges from the min or ax is of the ellipsoid to the major and hence at some stage of the rotation it passes through a value equ a l to the mean ax is At thi s stage t h e shadow and projecting cylinder are circul ar Th e nece s sary and s u fii c ie n t condi tion that r be the major or mi nor s emi-axi s of the sec t ion of the ellipsoid r O r 1 b y a plane passing through the center and perpendi cular to a is that a r and O r be coplanar . . , , . . . ' Let , , . r and Difi e re n t ia t e d 0 r w o r 1 r O a . r O 0, Furthermore r if r is to be a major or minor ax is of the section ; for is a max imum or a m in in u m and hence is perpendic ul ar to d r These three equations show t hat a r and O are all ortho gonal to the same vector d r Hence they are coplanar , r , . C onversely may Then dr be if . [a r O [a r O r] r] O . 0, chosen perpendicular d r . o r to 0 . t l heir common p ane . Q UA DR I C S URFA CE S 3 87 ' Hence r is a maximum or a m in in u m and is one of the prin c ip a l semi axes of the section perpendi cula r to a It is frequently an advanta ge to write the equati on of an elli psoid in the form . t in s t e a d o f This y Tz o r O o ma be done ; becaus e r o r o l, z l z . if jj ii kk z c ii + a ’ ii kk b 0 is a dyadic such that T is equal to O T may be re garded as a s u are root of O and written as O i B u t it must b e re memb ere d that there are o t her square roots of O for 3 q . . i i jj kk i i jj kk y y i thi s reason it is necessar to bear in mi nd that th e square root whi ch is meant b O 5 s that particular one whi ch has bee n denoted by T Th e equation of the e lli psoid may be wr tte n in the form Fo r i t T o o T o r L et r be the radius vect r of a unit the sphere r ’ o is ’ r ’ 1 l z . , phere S . Th e q e uation of VECTOR ANAL YSI S 38 8 If r i r it becomes ev dent that an ellipsoid ma y be transformed into a unit sphere by applying the operator T to each radi us vector r and vice ver s a the unit sphere may b e transformed into an elli psoid b y applying the inverse O per ator T I to each radi us vector r Furt hermore if a b c a re a s ste m of three conjugate radii vectors in an ell ipsoid ’ T o , , . y ’ ‘ . a . T -c , T -b Z a If for w o z a the moment a a z z z ’ , b o w o b o z T o b z ’ ’ 0 , b c z = c , , c o W T2 o C O 0 a denote respectivel I, Z = 0 . y T a T , b, ’ c r ’ Hence the th ee radii vectors a b c of the unit sphere into which three conjugate radii vectors in the ellipsoid are trans 1 T are mutually orthogonal They formed b y the operator form a right-handed or left-ha nded system of three mutuall y perpendic ul ar unit vectors Th e o re m : A n y elli psoid may b e transformed into an other elli psoid b y means of a homogeneous strain et the e uations of the ellipsoids be ’ ’ ’ , , “ . y . L q . . -O -r - r and z l, r ii r means of the strain O i the rad vectors of the fi rst ellipsoid are changed into the radii vec t ors r of a un it sphe e By r ’ r ’ = O4 r, o means of the strain T the radii vectors r of thi s unit sphere are transformed in like manner into the radii vectors r of the second elli psoid Hence b y the pro uct r is changed By ’ ’ i d . r= - . T i Ti 0 r . ( 1 9) 390 Iftwo VECTOR A NA L YSIS l y , L et the quadrics be r O r 1, and r T r 1 Let r -r a n d s = O s Then ri l confoca quadrics intersect the do so a t ght a n g es = O 1 " o s an d ’ . = Tr, = T r 1 “ q the uadr ics may be written in terms of s and s and s O ’ T s y -s = I ‘ I " s where b the confocal property —1 r Ifthe ’ as 1, ’ 1, , — I p r . quadrics inte sect at the condition for perpendicularity is that the normals O r and T r b e perpend cular That is i , . s B ut T 1 ‘ :c S In o s ’ s T ’ ‘ o l o ’ — s = s o T ‘ s T ( + 8 o —1 + 1) a s o 9c s , l s o ’ =1 T — s ‘ l o s ’ . like manner r — O ‘ I -S = ' x s o 2 s a T “ ’ o s o s -O s s I ’ " s ’ 1 (O — o s ’ (O “ I s ‘ l ml) -O T “ 1 ‘ 1 ) s a s ’ = O = s a o ‘ 1 -O ' s o s o ‘ 1 s ’ o ’ s — 1 . ’ . and the theo rem is proved If the parameter n b e allowe d to var from c th e c o to resultin g confocal qu a drics will consist of three fam ili es of which one is ellipsoids ; another hyp erb oloids of one sheet ; a n d the third h B y the foregoing yp erb oloids of two sheets y . o , , . Q UADR I C S URFA CE S y 391 y theorem each su rface of any one fam il cuts ever surface of t h e o t her two orthogo n ally Th e sur fa ces form a triply orthogonal system Th e lines of intersection of two famili es -sheeted and t h e fa mi ly of two-shee t ed say the family of one ( hyperb oloids) cut orthogonally t h e other family — the family of ellipsoids Th e points in which t wo ellipsoids are cut b y these lines are called corre s pondi ng points upon the two e llip It may b e Shown t ha t the ratios of the components of s o ids the radi us vector of a point to the ax es of the ellipsoid thr ough that point are the same for any two corresponding points Fo r let any elli psoid b e given by the dyadic . . . . . ii a kk : 1 $ 7 . neighb o rin g elli psoid in the family is represented by the dyadic Th e i i Z — a dn T “ I 2 — 6 O " dn 1 I dn . Inasmuch as O and T are homologous ( see Ex 8 p 3 30) dyadics they may b e treated as ordin a ry scalars in algeb ra Therefore if te rms of order higher tha n the fi rst in d n b e . , . . o m l t t e d’ Th e T O I dn . two neigh boring ellipsoids are then r r O 1, f ( O + I ci n - g 5 O dn) i Oi ‘ O d n) o r = r o r, -r , VECTOR A NAL YSIS 39 2 vectors f and r differ by a mul tiple of O r whi ch is perpendicular to the ellipsoid O Hence the te rmi ni of f and r are correspon di ng points for they lie upon one of th e lines which cut the famil y of ell ipsoids orthogonally Th e com :c and i o n e n t s of r and i in the direction i are r p Th e . , . at n r a: Th e Th e a . 1 1 Q 0 0 r L T: a: 2 rati o of these components dn is 2 a ax es of the ellipsoids in the direction are Va Thei ratio is i r V z — a 2 2 dn an d dn a In lik e manner _ b dn _ V 2 b 2 V0 2 3/ dfl 0 Hence t h e ratios of the components of the vectors 1 and r drawn to correspondin g points upon two neighboring e lli p fer at most by terms of the se cond order in al n s o ids o nl y di f from the ratios of the ax es of those ell ip s oi ds It foll ows immedi ately that the ratios of the components of th e vectors drawn to corresponding poin ts upon any two elli psoids sepa rated by a fi n i t e variation in the parame t er n onl differ at most b y terms of the rs t order in d n from the ratios of the ax es of the ellipsoids and hence must b e identical with them Thi s completes the demonstration " . fi y , , . . Th e P r op a ga t io n f o Ligh t in Cry s ta ls l electromagneti c equations of the ether or of any in fi ni t e i s o tropic medi um whi ch is transparent to e l e c tro ma g ne t ic waves may be wri t t en in the form 1 43 1 ] Th e To t re a t Th e g m m g m m fo llo wi n d i sc u s s io n u s t b e re a rd e d a s a t h e t h e s u bj e c t fro t h e s t a n d po in t o f p h y s i c s wo u ld b e a i l t ca o ut o f not p h ys i c a l pla c e he r e . . VE C TOR 3 94 D= A Then c os A NAL YSIS (m o r — t) n where A and m are cons t ant vectors and n a constant scalar Th e vi b ra ti ons ta ke place in represen t s a train of waves the direction A That is t h e wave is plane polari z ed Th e wave advances in t h e direct ion m Th e velocit v of that a d vance is th e quotient of n b y m the magni tude of t h e vector If this wave is an elec t romagnetic wave in t h e medium m considered it must sat isfy the two equa tions of that medium Sub s ti t ute t h e value of D in those equations Th e value of V D V V O D ma y b e D and V V O obtained most easily b y assum ing the direction i t o b e c o in c i dent wi t h m m r then reduces to m i r which is equal to mx Th e variables y and z no longer occur in D Hence . y , . . . , . . . o o , , . . . B = A V V V V Hence o o n 9D l z V O o a o V o — m — m -V V V — i -A — m2 D V m ( a: O 2 i i O A o x c os (m zc -O -A -D — -O -D — m e 2 o x m m O O o O o D -D . J D . Hence if t h e harmonic vib ration tions ( 4 ) of the medium D=m m ( c os m mm n 2 m ( -A sin ( n -r M oreover n s in m x D -O -D c os -D — -A = O . y D is to satis f the equa mm o O -D THE Th e be t r Th e PR OPA GA T ON OF L GHT I I CR YS TAL S IN 39 5 latter equation states at once tha t the vib ra tio n s m us t a n s ve rs e to the direction m of propagation of the waves former equation may be put in the form . m mm m o n 2 Introduce vector 8 is in the direction of ad vance m Th e magni t ude of s is the quoti ent of m b y n Thi s is the reciprocal of the velocit of the wave Th e vector 3 may therefore be called the wave-slowness O D= S s O D— s e D Th e . y . . . o This o . may als o be written as D= Dividing b y the scalar factor cos (m x -A) x s s s O -A — s s o -O -A (7) . evident tha t t h e wave slowness 3 depends not at all upon the phrase of the vib ration b u t only upon its directi on Th e motion of a wave not plane polariz ed may b e discussed b decomposing the wave into waves whi ch are plane polari z ed et a b e a vector drawn in t h e di recti on A of the displacement a n d let the magni tude of a be so determin ed It is y . L t hat Th e . -O -a a = l ( 8) . equation ( 7 ) t hen b ecomes reduced to the form O s s O s s s s a o a c O -a o z l (9) o ( 8) . These y are the equations b y which the discussion of the velocit or rat her t h e s lowness of propagation of a wave in different directions in a non-isotropic medi um may b e carried on s a O a = 3 s a = s a . o o o . VECTOR A NAL YSIS 3 96 Hence the wave slowness s due to a displacement in th e direction a is equal in magni tude ( b u t not in di rection) to the radius vector drawn in the elli psoid a O a 1 in that direction a x O s — a x s s O s -a ) -a S O s s 0 . o o z B ut the fi rst term contains a x -O s o s O a [a i o i tw ce and van shes s o . Hence . O 1 1 ( ) wave-slowness 8 therefore lies in a plane wi th t h e direction a of displacement and the normal O a drawn to the ellipsoid a O a 1 at the terminus of a Since 8 is p e rp e n dic ul a r to a and equal in magnitude to a it is evidentl com p l e t e ly determined e x cept as regards Sign when the d rection G iven the direction of displacement the line of a is known advance of the wave compatib le with the di splacement is com l e l determined the velocity of the a d ance is likewise v t e p y known Th e wave however may advance in either direction along that line B y reference to page 3 8 6 equation ( 1 1 ) is seen to b e t h e conditi on that a shall b e one of the principal ax es of th e ellipsoid formed by passing a plane through the e ll ipsoid perpendicular to 8 Hence for any given direction of advance there are two possib le lines of di splacement These are the principal ax es of the elli pse cut from t h e ellipsoid a O a 1 b y a pla ne passed through the center perpendicular to the line of ad vance To these statement s concerning the de t er mi n a t e n e s s of s when a is given and of a when s is given just such ex ceptions occur as are ob vi ous geometrically If a and O a are parallel 8 may have any di rection perpendi cul ar to a Th is happens when a is directed along one of the principal ax es of the ellipsoid If s is perpendicular to one of the c ircular sections of t h e ellipsoid a may have an d recti on in the p ane of the sectio n Th e . i y . , . . , . . . . . . l . yi 3 98 VECTOR ANA L YSIS L et j+ x Then i + y and k z 3 2 x 2 r 2 + y + z 2 . i the equation of the surface in Ca tesian co ord nate s is 2 equation in d rectly from i ( i Cartesian Th e l — O ‘ s s o y i coord nates ma be obta ned I+ s s y )8 _- O . determinant of this d adic is a z x 2 — b means of the relation forms By a 2 — 2 + 2 a y z s l y 2 2 3 y 2 — 3 a 2 + y b y 2 — 3 b 2 e z this assumes the 2 z + 8 2 s b z + s This equation appears to be of the six th degree It is how ever of only the fo u rth Th e terms of the six th order cancel out Th e vector s represents t h e wave-slowness Suppose that a plane wave polariz ed in the di rection a asses the orig n at a . . . p . i THE PR OPA GA T ON OF L GHT I I IN CR YS TAL S 399 certain instant of time wi th this slowness At the end of a unit of t ime it will have travelled in the di rection 3 a distance equ a l to the reciprocal of t h e ma gni t ude of s Th e plane will be in this position represented b y the vector 3 ( page . , . If s n i + v j+ w k plane at the ex piration of the unit time cuts o ffintercepts upon the ax es equal to the reciprocals of i t v w These quantities are therefore the plane co ordinates o f the plane They are connecte d with the co ordi na t es of the points in the plane b y the rel ation th e , , . . c = w z L + y If di fferent plane waves po la riz ed in all possible di fferent directions a be supposed to pass through the origin at the same instant they wi ll envelop a surface at the end of a unit of time This surface is known as the wave -surface Th e pe rp endic ul a r upon a tangen t plane of the wave-surface is the reciprocal of the slown ess a n d gives the velocit with which the wave travels in that direction Th e equation of the wave su rface in plane co ordi nates u v w is identical with the eq u a tion for the locus of the terminus of the slowness vector s Th e equati o n is . . y . , , . 3 2 b 2 1 where 3 n O This may be wri tten in any of the w forms given previously Th e su rface is known as Fre s n e l s Wa ve -Su rfa c e Th e equations in vector form are given on page 3 9 7 if the variab le vector 8 be regarded as determi ning a plane instead of a point In an isotropic medium the di rection of a ra y of li ght is perpendi cular to the wave-front It is the same as the di rection of the wave s ad vance Th e velocit of the ra 2 2 z 2 . ’ . . . . ’ . y y VECTOR ANA L YSIS 4 00 is equal to the velocity of the wave In a non isotropic medium this is n o lon ger true Th e ray does not travel per -front — that is in t h e dir ection of the to wave t h e a r n di c ul e p wa ve s advance A n d the velocity with whi ch t h e ray travels is greater t ha n the velocit of the wave In fact whereas the wave -front travels o ff a l ways tangent to the wave -surface the ray travels along the radi us vector drawn to the point of tan -planes -plane T h e nc of wave wave envelop e t h e th e g y wave-surface ; the termini of the rays are situated upon it Thus in the wave-sur face the radi us vector represents in mag n i t u de and direction the velocity of a ray and the p e rp e n dic ul ar upon the tangent plane represents in magnitude and direction the velocity of the wave If instead of the wave surface the surface which is the locus of the e x tremit of the wave slowness b e considered it is seen that the radius vector represents t h e slowness of the wave ; and the perpendicular upon the tangent plane t h e slowness of the ray et v be the velocit of the ray Then s v 1 because the ext rem ity of v lies in the plane denoted by 8 Moreover the condi tion t h at v b e the point of tangency gives d v per In like manner if s b e the slowness of the p e n di c u l a r to s ray and v the velocity of the wave 8 v 1 and the cond tion of tangency gives cl s perpendi cul ar to v Hence - . . , ’ y . . , , . . y . L y , ’ . ’ . ’ . ’ ’ ’ i . ’ , ’ . s and s v v y ’ v ma be e xpressed in terms of = a da = 2 s o 8 ds O 0 8 -a — s ds M ultiply b y a O -d a a = ¢ a 0 — S o O ’ d o and a , s, — 8 O o -a o = s = 0, 8 0 a s s ¢ O 0 s ’ o d v = 0, as follows a’ ds + s -s -O -d a O -d a . and take account of the relations 0and a a s Then s ‘ o o . . a s O and VECTOR ANA L YSI S 402 and the normal to the elli psoid the wave -slowness and the 1 on the other O ellipsoid It was seen t hat if s was n ormal to one of the cir c u la r sections of O the displacement a coul d take place in any direction in the plane of that section Fo r all directions in t hi s plane the wave -slowness had the sa me direction and the Hence t h e wave-surface has a singular s ame magnitude plane perpendi cular to s Thi s plane is tangent to th e surface along a curve instead of at a single point Hence if a wave travels in the di rection 8 the ray travels along the elements of t h e cone drawn from the cen t er of t h e wave-surface to this curve in whi ch the singular plane touches the s urface Th e t wo directions 3 whi ch are normal to the circ ul ar sections of O are called the p rim a ry Op tic a xe s These are t h e ax es of equal wave velocities but un equal ray velocities In like manner v b eing coplanar with a and O a , , — . . . . . . . . ’ O [ o a ’ v a ] last equation states that if a plane b e passed through 1 O the center of the ellipsoid perpendicular to v then a which is equal to O a will b e di rected along one of the prin Henc e if a ray is to take a de fi n i t e c ip a l a x es of the section di rection a may have one of t wo d rections It is more con ve n ie n t however to regard v as a vector determining a plane Th e fi rst equation O a v a 0 ] [ Th e ’ “ ’ , i . ’ . ’ . ’ stat es that a is the rad ius vector drawn in the ell ipsoid O to the point of tangency of one of the principal elements of t h e cylinder circumscribed ab out O parallel to v if b y a principal element is meant an element passing through the e xtremi ties of t h e major or mi nor ax es of orthogona l plane sections of th at cylinder Hence given the di rection v of the ray the t wo possib le directions of displacement are those radi i vectors ’ ’ . , VAR IA B LE DYADICS 4 03 of the ellipsoid whi ch lie in the principal planes of the c ylin der circumscrib ed ab out t h e elli psoid parallel to v If the cylinder is one of the two circula r cylinders which may b e circumscrib ed ab out O t h e direction of displ a cement may be any di rection in t h e plane passed through the center of t h e elli psoid and containing the common curve of tangency of t h e cylinder wi t h t h e elli psoid Th e ray-velocity for all these directions of di splacement has the same di rection and the same magni tude It is therefore a line drawn to one of the sin gular points of t h e wave -surface At this singular oint there are an in fi nite num b er of tangent planes envelop ing a cone Th e wave -velocity may be equal in magni tude and direction to the perpendicular drawn from the o ri gin to any of these planes Th e directions of the ax es of the two circular cylinders c irc um s c rip t ib l e ab out the ellipsoid O are t h e di rections of eq ua l ray-velocity b u t unequal wave velocity They are t h e radii drawn to the singular points of the wave surface and are called the s e c o n da ry op tic a xe s If a ray travels along one of the secondary optic ax es the wave planes travel along the elements of a cone ’ . . p . . . - . . . Va ri a ble Dy a dic s Th e Dife re n tia l . a nd In t egra l Ca lc u lu s Hi t herto the dyadi cs considered have b een constant Th e vectors which entered into their make up and t h e scalar coe ffi cients which occurred in the e x pansion in noni on form ha ve b een const ants Fo r t h e elements of the theory and for elementary applications the s e constant dyadi cs suffi ce Th e intro ducti on of variab le dyadi cs however leads to a s im p lifi c a tion and uni fi cation of the di fferent ia l and integral calculus of vectors and furthermore variab le dyadi cs b ecome a necessity in t h e more advanced applications for instance in the theory of the curvature of surfaces and in the dynamics of a rigid body one point of which i s fi x ed . . . , p , , , . VECTOR A NAL YSIS 4 04 Let W be a L ector function of position in space et r b e the vector drawn from a fi x ed origi n to an point in space y v x Ol Hence d ly — i + y j+ OW M 9 W= d r z . . k, SW SW sy x s ' z 9W 9W ly E Oz o expression enclosed in the braces is a dyadi c It thus appears that the di fferential W is a linear function of d r the differential change of position Th e antecedents are i k and the consequents the fi rst partial derivatives of W with re spect of x y z Th e ex pression is found in a manner precisel analogous to de l and will in fact be denoted by V W Th e . , , . , , j , , y . . VW— 9 , r W O 9 m W OW 9y Oz ' 1 ( ) 2 W d r VW ( ) This equation is like the one for the di fi e re n t ia l of a scalar func t ion V dV dr V V Then d . . : may be regarded as de fi ning nonion form V W b ecomes It VW — . . rr M . 9X OY a 9 x eg aX a z ” VW x sy . aY . If e xpanded into VECTO R A NAL YSIS 4 06 y an attempt were made t o appl the operator V s ym b o li cally to a scalar function V th re e times the result would be a sum of twenty -seven terms l ike If , 9 V 9 V 3 , ” x 9 This 3 9 x 9 y 9 z is a tria dic Three vectors are placed in jux taposition wi thout an Sign of mul tipli cation Such e x press ions will not b e discussed here In a similarmanner if the operator V b e applied twi c e to a vector function or o n c e to a dya d ic func tion of posit ion in space the result will be a triadic and hence ou t s ide the limit s set to t h e di scussion here Th e operators V x and V may however be appli e d to a d adic O to ield respectively a dyadi c and a vector y . . . , , . y y . 9 V o ¢ = Qy a: 9 O l o a x OO + 1 0 9y If where u, v , w are V v X O o vector functions of position in space O=V CW 9 9g Oz v r o u i + V a 17 . n Oz O 9 a 9 w + o v j+ V -w k , . OW 9 a: a v 9 W 9 y Oz ( 8) a similar manne the scalar operators ( 3 V) a n d ( V may be applied to O Th e resul t is in each case a dyad c In . i V) , 4 07 VA RI AB LE D YA DI CS a T 9 a) V ( . 9 V) O 2 9 O x 2 9 + 9 (17 O 2 9 y2 2 9 + 9 O 2 z ' operators a V and V V as applied to vector func tions are no longer necessaril y to be regarded as sin gle Oper a t o rs Th e individual steps ma be carried out by means of the dyadi c V W Th e y . . = V a W ( ) a -V) W = V (V -V W . when applied to a dyadi c the operators cannot b e inte r re t e d as made up o f two successive steps wi t hout ma k ing u se p of the triadi c V O Th e parent heses however may b e removed wi t hout danger of confusion just as they w ere removed in case of a vector function b efore the introduc t ion of the dyadi c Formul ae sim ilar to those upon page 1 76 may b e given for di fferentiatin g products in t h e case that the di fferentiation B ut . . l e a d t o dya di c s x . w -V w) V ( V o v — V w) = V v -Vw — -w + -v w + V v -V w -l- V w v V W v o o , v, -V w -V O, . V V x V x O : VV -O — V e tc . principle in these and all similar cases is that enun cis ted before nam ely : Th e operator V ma be t eated s m Th e , y r y VECTOR A NAL YSI S 4 08 as a vector Th e differentiations which it implies must b e carrie d out in turn upon each factor of a roduct to which it is appli ed Thus b o lic a lly p . . — v w . Hence A gain Hence A was seen ( rt 7 9) that if C denote a curve of ] whi ch t h e initial point is t and the fi na l point s r the li ne in te gra l of t h e deri vative of a scalar function taken along the curve is equal to the di fi e re n c e between the va ues of tha t function at r a nd r 1 48 It i . o l o . f a In like manner V V = V ( r) r f a r V W= f dr It o V V ( ro ) . W(t ) W ( r) o , W= 0 . may be well to note that the integrals f dr -V W and i i d are b y no means the same th ng V W is a dyad c Th e ve ctor al r cannot be placed arbitraril upon either si e of it . y . . VECTOR ANA L YSIS 41 0 surface integrals are taken over the complete bounding surface of t h e region t hroughout whi ch the volume integrals are taken N u merous form ul ae of integration by parts li ke those upon page 2 5 0 might b e added Th e reader wil l fi n d no di ffi cul t y in ob t ain ing them for himsel f Th e integrating operators ma also b e e x t ended to other cases To the potentials of scala r and vec t or functions the pote ntial P o t O of a dyadi c may b e added Th e N ewtonian of a vector function and the apla c ia n and Max wellian of dyadics may b e de fi ned Th e . y . . . L , , . . N ew r W rs W(x 2, 9 2, d 17 2 , r3 x $ 0132 r Q r Ma x O z , 3 a ) d 1) 12 ( x2 , y” dc 2 analytic theory of these integrals may be developed as Th e most natural way in which the demonst rations b efore may b e given is b y considering the vector function W as the sum of its components Th e . . , W= X i + Y j+ Z k q and the dyadi c O as e x pressed with the constant conse uents i k and variab le antecedents u v w or vic e ve rs a , j , These , , , , matters will b e left at thi s point Th e object of e u te ring upon th em at all was to in di cate the natural e x t ensions whi ch occur when variab le dyadi cs are considered These e x te nsions di ffer so sli ghtly from the simple cases which have . . THE CUR VA TURE OF S URFA CES 41 1 gone before that it is far better to leave the details to b e worked out or assumed from analogy whenever they may be needed rath er than to attempt to develop t hem in advance It is su fh cient merely to mention what the e x te ns ions a re and how the may be treate d y . . Th e C u rva f tu r e O Su rfa ces ] There are t wo difi e re n t methods of treati ng the cur In one the surface is e x pressed in pa ra va t u re of surf a ces met ic form by three equations . 33 = fr ( ) “a v , z = fs ( u y = f2 r = f u, ( r ) , . This is analogous to the method followed ( Art 5 7) in dealing wi t h curvat ure and torsion of curves and it is the meth od employe d b y Fehr in the b ook to whi ch reference wa s made In t h e second method t h e surfa ce is e x pressed b y a single equation connecting the variab les a y z — thus . . ' , F ( as, z y, ) , 0 . latte r method of treatm ents a ffords a simple applicat ion of M oreover the t h e differen t ial cal c ul us of variab le dyadi cs dyadi cs le a d naturally to th e most important resul t s connected wi t h the elementary t heory of surfaces et r be a ra di us vector drawn from an arb itrary fi x ed origin to a variab le point of t h e surface Th e increment d r lies in the surfac e or in the tangent plane drawn to the surface at t h e terminus of r Th e , . L . . . s dl ' O VF Z - O. ‘ Hence the derivative V F is collinear with the normal to th e surface M oreover inasmuch as F and the negative of F when 1 M uc h m , . o f wh a e s pe c ia ll y t ru e t fo ll o ws o ft h e t re a t m g p ra c t i c a lly fre e fro e n t o f e o d e t ic s A rt s is , . the u se 1 55-1 5 7 . of dy a di c s . Th is is VECTOR A NAL YSIS 41 2 y equated to z ero give the same geometric surface V F ma be considered as t h e normal upon either side of the surface In case the surface belongs to the famil de fi ned b y , y l F ( a y ' , , z . const ) . the norma V F lies upon that side upon which the consta nt inc reases et V F be epresente d b N the magn i tude of which may b e denote d b y N and let n be a uni t norma l drawn in the direction of N Then . L r y , . N e 2 = = N VF N o V F, y l q y is the vector drawn to an point in the tan gent p ane at the te rminus of r s — r and n are perpend c u la Conse uentl the equation of the tangent plane is If s i r , 6 — . = V F & fl li i and in like manner the equation of the normal ne s (s ) X r + kVF r s = VF O, q where k is a variable parameter These e ua tions ma y be translated into Cartesian form and give th e famil iar esul t s Th e variation d n of the un i t normal to a su rface plays an important part in the theor of cu vatu re d a is perpendic ul ar to 11 because 11 is a unit vector . y r . r . . VECTOR A NA LYSIS 41 4 the surface form is possib le ( Art It . r to educe 1 1 6) . O to the 5 ( ) j y where i and are two perpendi cula r unit vectors l ing in the tangent plane and a and b are positive or negative scalars ’ ’ . a vectors i of the surface n = d r o j and the sc ( a ’ ’ fi + b ars a b vary from point t o point Th e dya dic O is variab le Th e conic r O r 1 is called the in di catri x of the surface at the point in question If this conic is an ellipse that is if a and b have the same Sign the surface is convex at the point ; b u t if t h e conic is an hyperb ola that is if a and b ha ve opposite signs the surface is concavo-convex Th e curve r O r 1 may b e regarded as approx imately equal to the intersection of t h e surface with a plane drawn parallel to the tangent plane and near to it If r O r be set equal to z ero the resul t is a pair of s t ra ight lines These are the a s ym p totes of the conic If they are real t h e coni c is an hyperb ola ; if imaginary an ellipse Two directions on the surface which are parall el to conjugat e diameters of the conic are called con j u gate directions Th e directions on the sur fac e which coin cide with the directions of the principa l ax es i of the indi c atrix are known as the principal di rections They are a spe c i a l case of conjugate directions Th e directions upon the surface which coincide with the direc t ions of the asymptotes of the indi catrix are known as as ymptotic directions In case the surface is conve x t h e indicatri x is an ellipse and th e asymptotic direc t ions are imagina ry In Special ca s es the dyadi c O may b e such that the c o e f fi ciente a and b are equal O may then be reduced to the form ’ Th e , ’ al , . . . , , , , , . . . . , . . ’ , . . . , . . ¢ = a a f+ rn j ’ THE CUR VA TURE OF S URFA CE S j 41 5 in an in fi nite num b er of ways Th e directions i and may b e any two perpendi cular di rections Th e indi catrix b ecomes a circle A n y pair of perpendicular di ameters of thi s circle give principal directions upon the surface Such a poin t is called an u m bili c Th e surface in the neighb orhood of an um b ili c is convex Th e asymp t otic directions are imaginary In another special case the dyadi c O becomes lin ear and redu Cib le t o the { 0 1 m a i i ’ . ’ . . . . . . f f . indicatrix consists of a pair of parallel lines perpendicular to i Such a point is call ed a parab olic point of the surface Th e further di scussion of these and other special cases will b e omitt ed Th e quadric surfaces afford e x ample s of t h e vari ous kinds of points Th e ellipsoid and the hyperb oloid of two sheets are conve x Th e indicatri x of point s upon them is an ellip s e Th e in Th e hyperb oloid of one sheet is concavo -convex Th e indicatri x di c a t ri x of points upon it is an hyperb o a of any point upon a sphere is a circle Th e points are all Th e indicat rix of any point upon a cone or cyli nder u m b ilic s is a pair of parall el lines Th e points are parab olic A sur face in general may have upon it points of a ll ty pes — ell iptic hyperb olic parab olic and um b ilical A li ne of princ ipal curvature upon a surface is a 1 52 ] curve whi ch h a s at each point the dire c tion of one of t h e prin c ip a l ax es of the indi catri x Th e direction of the curve at a point is always one of the principal direc t ions on the surface at that point Through any given point upon a surface two per n di c u l a r lines of prin c ip a l curvature pass e hus the lines T p of curvat ure divide the surfac e into a system of in fi ni te s i mal rectangles A n asympt o t ic line upon a surface is a curve which has at each poin t the direc t ion of the asym ptotes of the indicatri x Th e direc t ion of the curve at a point is al ways one of the as mptotic directions upon the surface Th rough Th e ’ . . . . . . l . . . . . . , . , , . . . . . y . VECTOR A NA L YSIS 41 6 p given point of a surface t wo asymp t otic lines ass These lines are imaginary if the su rface is convex Even when real they do n o t in general intersect at righ t an gles Th e angle between t h e two a s ymptotic lin es at any point is b isected by the lines of curvature which pass through tha t point Th e necessary and su ffi cient condition that a curve upon a surface b e a line of principal curvature is that as one advances along that curve the increment of d n the uni t normal to the s u rface is parallel to the line of advance Fo r any . . . . , , . d n = O -d r -d r = ( a dr a: y Then evi dently d n and d r are parallel when and onl when ’ d is parallel to i or j Th e state ment is therefore proved It is frequen t ly taken as the de fi n i tion of lines of curvatu e Th e di fferent ial equation of a li ne of curvature is r ’ r . dr = 0 a . . ( 6) . A nother method of statement is that the nor mal to the surface the increment d n of t h e normal and the element d r of the surface li e in one plane when and only when t h e element at r is an element of a lin e of principal curvature Th e differential equa tion t hen b ecomes , , . [ y 11 a d r] = 0 n ( 7) . necessar and su ffi cient condi tion that a cur ve upon a surface be an asym ptoti c l ine is that as one advances a l ong t hat curve the increment of the unit norm al to the surface is perpendic ul ar to the line of ad vance Fo r Th e , . dn = d r da If t hen d n -d r z dr o O O -d r . is z ero d r O d r is z ero Hence d r is an asymptotic direction Th e state ment is t herefore proved It dr . . . VECTOR A NAL YSIS 41 8 0: Hence C C d a a cos ( i a cos ( i 2 2 a t ’ dr + b d r) , ’ d r) , a e dr r 2 ) -d r b cos (j d r) , b sin ( i d r) ’ 2 , 2 ’ , 1 0 ( ) . r interpretation of this formula for the cu vature of a normal section is as follows : When the plane p turns about the normal to the surface from i to the curvature C of the plane section varies from the value a when the plane passes ’ through the principal direction i to t h e val ue b when it passes through the other principa l d rection Th e val ues of the curvature have algeb raically a max imum and minimum in the directions of the principal li nes of curvature If a and b ha ve unl ike signs that is if the surface is concavo -conve x at P there ex ist two di rections for whi ch the curvature of a normal section vanishes These are the as mptotic directions Th e sum of the curvatures in two normal secti ons at ri ght angles to one another is const ant and independent of the actual position of those secti ons Fo r the curvatu e in one section is Th e j ’ ’ , i , j ’ . . , , y , . . r . C1 and in a (308 2 (i ’ , d r) b sin ( i 2 ’ , d r) , section at right angles to th is C2 a sin ( i 2 ’ d r) , b cos ( i 2 ’ , d r) . Hence which proves the statement It is easy to show that t h e invariant O s is e ua to the pro z duc t of th e curvatures a and b of the lines of principal c urv ature q l . . q Hence the e uation a x2 O, a: b O2 5 THE CUR VA TURE OF S URFA CES 41 9 is the quadratic equation whi ch determines the principal c u rv atures a and b at any poin t of the surface B y means of this equation the scalar quantit es a and b may b e found in terms of F ( x y z ) i , . . , n n ) — I ( VVF o n u ) N VVF ( n u 2 — VVF+ n n VVF o -n n -V V F ) S -V V F V V F ( )S Hence ( = ( n n o n n . v V V F)S n u N v r v ( ( n n -V V F ) S H ence V . r m v z VF o V F, VVF= n n : -V F V Th ese n n VF VF z N 3 VF V VVF n o u . VVF VF X F J y y i ex p e ssions ma be written out in Cartesian co ord na tes but the are e x tremel long Th e Cartesian ex pressions for O2 5 are even long er Th e vector e x p ession may be o b tained y r . . as , fo ll o ws : (I “ 11 1 02 V V F )2 ( I ( — un g ) 2 7 A I ( Hence VF . F V V )2 ( N G iven 4 y — nn . )2 VF VF VF : N ( V V F) 2 . 4 L an curve upon a surface et t be a unit tangent to the curve n a unit norma l to the su face and m a . , r VECTOR A NAL YSIS 4 20 vector de fi ned as n x t Th e three vectors n t m constitute an i k system Th e vecto r t is para l lel to the element at Hence the cond tion for a line of curvature becomes , j , i r , , . . Hence t x d n = 0 m n z o d . o -a n n m o a n . + n o a m . = 0 . -d m = 0 . t x d m = 0, 1 6 ( ) d m x dn = 0 . p increments of m and of n and of r are all arallel in case of a line of principa curvature A geodetic lin e upon a surface is a curve whose osc ul ating plane at each point is perpendicular to t h e surface That the geodetic li ne is the shorte st line whi ch can be drawn b e t ween t wo points upon a surface may be seen from the follo wi ng considerations of mechanics et the surface be smooth and let a smooth elastic string which is constrained to lie in the surface b e st retched between any two points of it Th e string actin g under its own tensions will take a position of e q u ili b rium along the shortest curve which can be drawn upon the surface between the two given points Inasmuch as th e string rs at rest upon the surfac e the normal rea ctions of the s u rface must lie in the osculating plane of the curve Hence t hat plane is normal to the surface at ever point of the curve and the curve it self is a geodetic line Th e vectors t and d t lie in t h e osculati ng pla ne and deter mine that plane In case the curve is a geodetic the normal to the osculating plane lies in the surface and consequentl is perpendicul ar to the norm al n Hence Th e l . . L . . . y . . . , . y 422 VECTOR A NAL YSIS i given surface is la d o ff Th e te rmin us P of this normal lies upon the surface of a sphere If the normals to a sur face at all points P o f a curve are thus const ructed from the same origin the points P will trace a curve upon the surfa ce of a unit Sphere This curve is called the spherical image of the given curve In like manner a whole re gion T of the surface ma Th e region T upon b e mapped upon a region T the sphere the sphere has b een called t h e ho do g a m of the region T upon the su rface If d r b e an element of a re upon the surface the correspondin g element upon the uni t sphere is ’ . . , ’ y . . ’ ’ . r . T dn dr o . be an element of area upon the surface the corre d element upon the sphere is a where s p o n din A r t ( g If d a , ’ . d a = O2 ’ O= O2 = a Hence b i x d a da . a ’ ’ o a ’ = j ’ j i x b n n ’ -d a a b n u . . tio of an element of surface at a point P to the area of its h o do gra m is equal to the product of the principa l radii of curvat ure at P or to the reciprocal of the product of the prin c ip a l c urvatures a t P It wa s seen t hat the measure of t urning to t h e right or left is m d t If then C is any c u rve drawn upon a s u rface the total amount of turn ing in advancing along the curve is the inte gral Th e ra . . . m dt . 2 ( 0) any closed curve thi s integral ma y b e evaluated in a manner analogous to that employed ( page 1 9 0) in the proof of Stokes s theorem C onsider two curves C and C near Fo r ’ ’ . C UR VA TH E TURE RFA CES 42 3 SU OF to ge t her Th e variation which the inte gral u ndergoes when t h e curve of inte gra tion is changed from C to C is . ’ B f m -d t f 8 z m ( d 8 f -d t = m (m -8 t) f 8m o f -8 t = dm dt + m f dm — -d t em d t) o o f m + . 8 dt d 8t o 8t + f (m d integral of the perfect di fferential d ( m when taken around a closed curve Hence St ) Th e -8 t) . vanis hes . f S m -d t z f dm dt o — f dm dt o . Th e ide m fa c t o r is Sm o dt = 8m o l o d t = 8m for t cl t and 8 m m vanish be effected upon the term at m . f S m o f dt= B y di fi e re n t ia t in g 8 m ( o n the relations o n n dt o similar transformation ma Then 8t A . -d t n m — n dm o u 0 and — m n n -B t ) t . 0 it is seen that Hence S 8 f m f m -d t f = m ( y -3 n t -d n — o f du t o -Sn ) dn x d n . VECTOR ANAL YSIS 424 x ferential 8 n Th e di f the h o do gra m r represents the element of a ea upon the unit sphere Th e inte gra dn l . f n o in dn x d n epresents the to ta l area of the h o do gra m of the s tri p of surface which li es between the curves C and C et t h e curve C start at a point upon the surface and spread out t o any desired siz e Th e total amount of turni ng whi ch is re qu red in making an infi nite simal circuit about the oint is 2 Th e to tal variation in the integral s r ’ . i i ff m dt o f m z da n B p . d ut -d t ’ H — 2 7r. , if H denote the total area of the h o do gra m H ence f m -d t H = 2 H + = 2 7r — L H . , 7r f m o dt = 2 rn area of the h o do gra m of the region enclosed closed curve plus the total amount of turning along that curve is equal to 2 7r If the surface in ques t ion is convex t h e area upon the sphere wil l appear positive when the curve upon the surface is so describ ed that the enclosed area appears positive If however the surfac e is concavo-convex the area upon the s phere will appear negative Thi s matter of the sign of the h o do gra m must be taken into accoun t in the state ment made above Th e . " . , , . . VECTOR A NAL YSIS 4 26 Vibra tio n s a rm o n ic H a nd B ive c to rs differential equation of rectilinear harmonic Th e motion is — n z z . d y l integral of this equation may be reduce b a suitab e choice of the constants to the form Th e A mz This sin t n . represents a vibration back and forth along the X-axis ab out t h e point x 0 et the displacement be denoted b Th e equation may b e written D in place of :c L . y . D= i A C onsider This D i A s in n sin n t t cos . m x . y is a displacement not merel near the point but along the entire ax is of a . At poi nts 2 h x rr , m 0 se where is a positive or negative integer the di splacement is at a ll Th e equation represents a stationar t imes equal to z ero wave with nodes at these points A t points midwa between these t h e wave h a s points of max imum vib ration If th e equat ion b e regarded as in three variables x y z it re p re sen t s a plane wave the plane of which is perpendi cular to the ax is of the variab le x Th e dis placement given by the equation k y , y . . . , , . D1 = i A l cos ( m ac — n (1) t) is likewise a plane wave perpendicular to the axis of a but not stationar Th e vib ration is harmonic and advances along the direction i with a velocit equal to the uotient of y . y q HAR M ON I C n by m length . VIB RA TI ONS be the velocity ; If c A N D B I VE C TORS 42 7 the pe ri o d ; and l the wav e p , 2 W m Th e a displacement cos ( m a — n t ) differs from D1 in the partic ul ar that the displa cement ta kes pl a ce in the direction not in the di rection i Th e wave as b efore procee ds in the di rec t ion of x with the same veloci t y This vibration is t rans verse instead of longitudi nal By a simple e x te nsion it is seen that D2 = j A 2 j . , . . D=A c os (m ax— n t) is a displacement in the di rection A Th e wave advances along the dir ection of c Hence the vib ration is ob lique to the wave -front A still more general form may be ob tained b substituting m r for m 93 Then . . y . . . D= A This c os (m o r is a displacement in the direction A Th e maximum amoun t of that d splacement is the magni tude of A Th e wave advances in t h e direction In ob lique to the displace ment ; the velocity period and wave-length are as b efore Elliptic ha r So much for rectil inear harmonic motion monic motion may be de fi ned b y the equation ( p i . . , . , . . cl 2 d i Th e r z n 2 r. general integral is ob tained a s r = A c os n t + B s in n t . discuss ion of waves may b e carried thr ough as pre vi o us ly Th e general wave of ellipti c harmonic motio n adva ncing in the recti on m is seen to be Th e . di VECTOR ANA L YSIS 42 8 D = A c os (I D — n dt (m o A s in r — n (m o t) — B s in (m o r — n . — n r y 4 ( ) t) t) i ( 5) y is the velocit of the displ a ced point at an moment in t h e ellipse in which it vib rates This is of course entirely differ ent from the velocity of the wave An interesting result is obtained by subtracting the dis placement from the velocity multipli ed b the imaginar unit V 1 and d vided by n . . y i — dt n + V — 1 y . A cos ( m -r — t) — B n sin ( m -r — n t) 6 ( ) i n A s j (m 9 “ d. n -r — n t ) + B s in V— + l n (m V: . -r r— u . e xpression here obtained as far as its form is concern ed is an imaginar vector It is the sum of two real vectors of whi ch one has be en m ul tiplied by the imaginar sc a lar V 1 Such a vecto r is called a bi ve c to r or i m a gin a ry vector Th e ordinar imaginar scalars may be called bisc a la rs Th e use of b ive c t o rs is foun d ver convenient in the discussion of el lipti c ha rmonic motion Indeed any undamped ellip t ic har m o n rc plane wave may be represented as above by the pro duct of a b ivector and an e xponenti al fac tor Th e real part of the product gives the displ a cement of any point and the pure imaginary part gives the velocity of di splacement reversed in Sign and divi ded b y n 1 59 ] fe s from that of Th e analyti c theory of b ive c t o rs di f real vecto rs very m uch as the an a lyt ic theory of b is c a l a rs differs from t hat of real scalars It is unnecessar to have an dist nguis hi ng cha a cter for b ive c to rs jus t as it is need Th e y y , , y . y . . y . . . . y i r . r y VECTOR A NAL YSIS 43 0 r 1 ‘ o ( rl s X S= ( 1 ' 1 x 8 — l s a r ' 5 1 X ' 3 i 2) ( 1 8 1 r 2 + 2 ” s 1 r X z 3 1 ) ) or b is c a la rs are said to b e conjugate when their real parts are equal and their pur e imaginar parts Th e conjugate of a real scalar or vector difi e r only in Sign is equal to the scalar or vector its elf Th e conjugate of an sort of product of b ive c t o rs and b is c a l a rs is equal to the pro duct of the conjugates taken in the same order A similar statement may be made concerning sums and di fferences Two b ive c t o rs y y . . . . ) ( fr + i r 2 ) 1 r1 = 2 i 2 x r1 , i r e “ r ' l 1 " l ' 1 y i If the bivector r rl i r2 be multiplied by a root of unit or cyclic factor as it is frequently called that is by an imag nary sc al ar of the form , where a 2 2 O ' 12 are unaltered Thus if I X ' 1 1a 1 1, : the conjugate is multiplied by products 3 , i b, a 1 2 4 " e and hence ra rl th e four " m u ltiplying the bivector r by such a fa cto r by I I I . HA RIlI ON I C VIB RA TI ONS AND B I VE C TOR S y 4 31 closer e xami nation of t h e e ffect of multipl ing a bivector b y a cyclic fac t or ields inte resting and import ant geometric results et A . y L ( cos i sin q) ( i r 1 cos q r 2 sin q, r, cos q r 1 sin q r 1 B y reference du c e d in the i q 8 ( ) . to Art 1 29 it will be seen that the change pro real and imaginary vector parts of a b ive c to r by m u ltiplication with a cyclic fac t or is precisely the same as would be produced upon those vectors by a cyclic dyadi c . , , O= a a ’ + c os — q sin g (c b ’ — b c ’ ) used as a prefac tor b and c are supposed to be t wo vectors collinear respectively with r1 and r, a is any vector not in their plane Consider the elli pse of whi ch rl and r2 are a pa ir of conjugate semi -diameters It then appears that rl and r2 are a lso a pair of conjugate semi-di ameters of that ellips e They are rotated in the ellipse from r1 toward r2 b y a sector of whi ch the area is to the area of the whole ellipse as q is t o 2 Such a change of position has been called an elliptic rotation through the sector q Th e ellip s e of which r and r are a pair of conjugate semi l 2 diameters is called t h e dire c tio n a l e llip s e of th e bi ve c to r r When the b ivector has a real direction the di rectional ellipse reduces to a right line in that direction When the b ivec t or has a comple x di rection the ellipse is a true ellipse Th e angular direction from the real part r1 to the complex part r2 is considered as the positive direction in t h e dir ec t ional el lipse and m u st al ways b e known If th e real and imagi nary parts of a b ivector turn in the positive dir ection in the ellipse they are said to be advanced if in the negative dirc o tion they are said to be retarded Hence mu ltip lic a tio n of a . . . ’ . ’ , . . . . . , . . VE C TOR ANAL YSIS 4 32 bi ve c to r by a a s e c to r e q u a c yc lic re t a rds f a c to r l t o th e a n l e g i t i n i ts direc tio n a l f the c y c li c fa c t o r o el lip se by y . is always possib le to multiply a bivector b such a cycli c factor that the real and imaginar parts become coincident with the ax es of the e llipse and are pe rp endi cular It y . To accomplish the reducti on proceed as follo ws r Ifa o o Form r b = 0, r -r — b r o tau With this value of b) . + i b, a r o b 2 q a i l p the ax es of the d rectiona elli se q a re given by the equation a + i b = ( cos q — y i s in q) r . case the real and imaginar par t s a and b of a b ivector are equal in magnitude and perpendic ul ar in direction b oth a and b in the e x pression for r r vanish Hence the angle A is indeterminate rectional ellipse i s a circle T h di e q b ivector whose di rec t ional ellipse is a circle is called a c i rc u l a r bi ve c to r Th e necessar and s u ffi cient con di tion that a non-vanishing bivector r be circular is In . . . y . r r condition r is not su ffi cient Th e r to r o r circular . r which for real vecto rs implies ens u re the vanishing of a bivector O, . r VECTOR ANA L YSIS 434 Th e fi rst equation shows that r2 and 3 2 are perpendicular a n d hence s 1 and 8 2 are perpen di cular M oreover the second shows t hat the angular direc t ions from r1 t o r2 and from 3 1 t o t h e a x es of the direc t ional ellipses are the same and that 8 2 of r and s are propor t ional Hence t h e conditions for perpendiculari t y of t wo b ive c t o rs whose planes coincide are that their direc t ional el l ipses are similar the angular direction in b oth is t h e s ame and t h e 1 major ax es of the ellipses are perpendicular Ifb o t h vectors have real directions the condi t ions degenerate into the per T h of hose directions n di c u l a rit t e conditions t herefore e p y hold for real as well as for imaginar vectors et r and s b e t wo perpendicular b ive c t o rs the planes of which do not coincide Resolve r1 and r2 ea c h in t o t wo com pone nte respectively parallel and perpendicular t o the plane of s Th e components perpendicular to that plane contrib ute nothing to t h e value of r 3 Hence the components of r1 and r2 parallel to t h e plane of 8 form a b ivector r whi c h is perpendic u lar to 8 To thi s b ivector and s t h e conditions stated ab ove apply Th e di rectional ellips e of t h e b ivector r is eviden t ly the projection of the directional ellipse of r upon the plane of 3 Hence if two b ive c t o rs are perpendicular the di rectional ellipse of either b ivector and the dire c tional ellips e of the other projected upon t h e plane of that one are similar have the same angular direction and have their major ax es per . , , . , , . y . L . . . . ’ . ’ . . , , , n d i u l r e c a p . C onsider a bivector of n th e 9 ( ) A where A and m are b ive c t o rs and position vec t or of a point in space . 1 t he m It s h o ul d be sa e as t he type is a b is c a la r r is the It is t herefore to b e con n . mm g d t h a t t h e c o n d i t i o n o f p e rp e n di c u l a ri t y o f a j o r a x e s is n o t c o n di t i o n o fp e rp e n di c ul a ri t y o f re a l p a rt s a n d i a in a ry p a rt s no te . HARM ONI C VI B RA TI ONS AND B I VE C TOR S 4 35 as real t is the scalar variable time and be considered as real et si de re d . . L A AI i A m m1 i m2 , n I) D A ( I A ( I = n i Aa ) i A2 ) 1 e e i ' ( mx o - m e ' n r g, + t mg m e r a g r — n1 ua t rd —mt) has been seen b efore the factor ( AI + i A2 ) represents a t rain of plane waves of ell ip t ic harmonic vi b ra tions Th e vib rations ta ke place in t h e plane of AI and A2 in an ellips e of which AI and A2 are conjuga t e semi di am Th e di s placement of t h e vi b ra t ing poin t from t h e e t e rs cente r of t h e ellipse is given b y the real part of t h e fac t or Th e veloci t y of t h e poin t reversed in di rec t ion and divided b y u 1 is given b y the pure imaginary par t Th e wave a d vances in t h e direction m l Th e other fac t ors in t h e e xpre s m sion are dampers Th e factor e is a damper in t h e di rec t ion m2 A s the wav e pro c eeds in t h e direc t ion 1112 it dies a way Th e fac t or e is a damper in time If n 2 is negative the wave dies away as time goes on If n 2 is posi tive t h e wave incre a ses in energy as time increas e s Th e presence ( for un limi t ed t ime ) of any such factor in an e x pre s sion which represents an ac t ual vi b ra t ion is c learly inad missib le It con t radic t s t h e law of cons erva t ion of energy In any physical vi b ra t ion of a con s erva t ive s y stem n 2 is n e c e s s a ril y nega t ive or z ero Th e general e x pression ( 9 ) t herefore represents a train of plane waves of ellipt ic harmonic vib ra t ions da mped in a de fin ite direc t ion and in t ime Two such waves may b e com pounded b y addi ng t h e b ive c t o rs which represen t t hem If the e x ponent m r n t is t h e same for b o t h the resul t ing train of waves advances in the same direction and has t h e As , . , - . . . . ‘ " r . . " " . . . . . . . . . VECTOR A NAL YSIS 43 6 same period and wave -length as the individual waves vib rations however take place In a di fferent ellipse waves are — -— , , A g . um the resultant is r n t) a nd B) (A e B g flm r a —nc If the a t) t un Th e . . comb i n i ng two t rains of waves which advance in opposite direc t ions b u t wh ich are in other respects equal a system of sta t ionary waves is ob tained By . A Ae — mz o r e — mc . f 6 { ml e 011 1 3 o . r — n t) A e — mg ' r e i ( — m1 -— (m 1 r r . t) n ) r e - mz d ' 6 theory of b ive c t o rs and their applications will not be carried furt her Th e ob ject in entering a t all upon t his very shor t and condensed dis cus sion of b ive c t o rs was fi rs t to show t h e reader h o w t h e s imple idea of a di e c ti o n has to give way to t h e more complicat ed b u t no less useful idea of a dire c tio n a l e l li s e when t h e generali z ation from real to imagi n ar ve c tors p is made and second t o set fort h the manner in which a single b ivector D may be employed to represent a train of plane waves of elliptic harmonic vi b ra t ions This application of b i vec t ors may b e used to gi ve the Theory of ight a wonderfully 1 simple and elegant trea t ment Th e . r y , L . m dL by . P ro fe s s o r Gibb s in h is c o u rs e o f l e c t u re s o n Th E l t m g t i Th o y of ight d e l i ve re d bi a n u a ll y a t Ya l e Un i ve rs i t y B ive c t o s we re n o t u s e d in t h e s e c o n d p a rt o f t h i s c h a p t e b e c a u s e in t h e O pi i o n o f t h e p re s e n t a u t h o r t h e y p o s s e s s n o e s s e n t i a l a dva n t a e o ve r re a l ve c t o rs u n t il t he o re a d va n c e d p a rt s o f t h e t h e o y ro t a t i o n o f t h e pl a n e o f p o la ri a t i o n by a n e t s a n d c ry s t a ls t o t a l a n d a re re a c h e d e t a ll ic re fle c t i o n e t c 1 Su c h e us e o e c ro f bive c t o rs is a ne c e r a e n , . g r m mg n r, , m r z , , . , .