MATHEMATICAL PHYSICS EUGENE BUTKOV St. John’s University, New York tte: ‘or Shug. rt ADDISON-WESLEY PUBLISHING COMPANY Reasling, Massachusetts » Menlo Park, California - London » Sydnay - ManilaSS ¥ g g 2 8 = Dep Stupe® FIRST PRINTING 1973 A complete and unabridged reprint of the original American textbook, this World Student Series edition may be sold only in those countries to which it is con- signed by Addison-Wesley or its authorized trade distributors. It may not be re-exported from the country to which it has been consigned, and it may not be sold in the United States'of America or its possessions. Copyright © 1968 by Addison-Wesley Publishing Company, Inc. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, with- ‘out the prior written permission of the publisher. Original edition published in the United: ‘States of America. Published simultaneously in Canada. Philippines copyright 1968 Library of Congress Catalog Card Number: 68- (1391 | PREFACE During the past decade we have witnessed a remarkable increase iri the number of students secking higher education as well as the development of many new colleges and universities. The inevitable nonuniformity of conditions present in different institutions necessitates considerable variety in purpose, general approach, and the level of instruction in any given discipline. This has naturally contributed to the proliferation of texts on almost any topic, and the subject of mathematical pliysies is no exception. There is a number of texts in this field, and some of them are undoubtedly of outstanding quality. Nevertheless, many teachers often feel that none of the existing texts is properly suited, for one reason or another, for their particular courses. More important, students sometimes complain that they have difficulties studying the subject from texts of unquestionable m This is not as surprising as.it sounds: Some texts have an encyclopedic character, with the material arranged in a different order from the way it is usually taught; others become too much involved in complex mathematical analysis, preempting the available space from practical examples: still others cover a very wide variety of topics with utmost brevity, leaving the student to struggle with a number of difficult questions of theoretical nature. ‘True enough, a well-prepared and bright student should be able to find his way through most of such difficulties. A less-gifted student may, however, find it very difficult to grasp and absorb the multitude of new concepts strewn across an ad- vanced text. ‘Under these ciscumstances, it seems desirable to give more stress to the peda- gogical side ofa text to make it more readable to the student and more suitable for independent study. Hopefully, the present work represents a step in this direction. It has several features designed to conform to the path an average student may conceivably follow in acquiring the knowledge of the subject. First, the inductive approach is used in each chapter throughout the book. Fol- lowing the fundamentais of modern physics, the text is almost entirely devoted to linear problems, but the unifying concepts of linear space are fully developed rather Jate in the book after the student is exposed to a number of practical mathematical techniques. Also, almost every chapter starts with an example or discussion of elementary nature, with subject matter that is probably familiar to the reader. The introduction of new concepts is made against a familiar background and is later extended to more sophisticated situations. A typical example is Chapter °, where the basic aspects of partial differential equations are illustrated using the “elementary functions” exclusively. Another facet of this trend is the repeated use of the harmonic oscillator and the stretched string as physical models: no vvi PREFACE attempt is made to solve as many problems for the student as possible, but rather to show how various methods can be used to the same end within a familiar phys- ical context. In the process of learning, students inevitably pose a number of questions necessary to clarify the material under scrutiny, While most of these questions naturally belong to classroom discussion, it is certainly beneficial to attempt to anticipate some of them in a text. The Remarks and many footnotes are designed to contribute to this goal. The author hopes they answer some questions in the mind of the student as well as suggest some new ones, stimulating an interest in further inquiry. A number of cross-teferences serves a similar purpose, inviting the reader to make multiple use of various sections of the book. The absence of numbered formulas is intentional: if the student bothers to look into the indicated section or page, he should not simply check that the quoted formula “is indeed there,” but, rather, glance through the text and recall its origin and meaning. The question of mathematical rigor is quite important in the subject treated here, although it is sometimes controversial. It is the author's opinion that a theo- retical physicist should know where he stands, whether he is proving his own deduc- tions, quoting somebody else's proof, or just offering a reasonable conjecture. Consequently, he should be trained in this direction, and the texts should be written in this spirit. On the other hand, it would be unwise to overload every student with mathematics for two main reasons: first, because of the limitations of time in the classroom and the space in a text, and second, because physicists are apt to change their mathematical postulates as soon as experimental physics lends support to such suggestions. The reader can find examples of the latter philosophy in Chapters 4 and 6 of the text. Whether the author was able to follow these principles is left to the judgment of users of this book. Each chapter is supplied with its share of problems proportional to the time presumed to be allotted to its study. The student may find some of the problems rather difficult since they require more comprehension of the material rather than sheer technique. To balance this, a variety of hints and explanations are often supplied. Answers are not given because many problems contain the answer in their formulation ; the remaining ones may be used to test the ability of the student for independent work. The exercises within the text can be used as problems to test the students’ manipulative skills. For many of the methods of instruction of mathematical physics presented in this book, the author is indebted to his own teachers at the University of British Columbia and McGill University. The encouragement of his colleagues and students at St. John’s University and Hunter College of the City University of New York is greatly appreciated. Also, the author wishes to thank Mrs. Ludmilla Verenicin and Miss Anne Marie Nowom for their help in the preparation of the manuscript. Palo Alto, Calif. EB. August 1966 Chapter 1 1d 12 13 14 1s 16 17 18 19 Chapter 2 21 22 23 24 25 2.6 27 28 29 2.10 211 212° 2413 2.14 215 Chapter 3 3.1 32 33 34 35 36 CONTENTS Vectors, Matrices, and Coordinates Introduction . 1 Vectors in Cartesian Coordinate Systems. 1 Changes of Axes. Rotation Matrices toe eee 4 Repeated Rotations, Matrix Mulliplication. 2 2... kk 8 Skew Cartesian Systems. Matrices in General woe ee OH Scalarand Vector Fields © 2 2 2... ew. Vector Fieldsin Plane. 2 2 2 2 we wwe. Vector FieldsinSpace. 2 2 2 2 2. kk Curvilinear Coordinates 6 we Functions of a Complex Variable Complex Numbers... wee ee AF Basic Algebra and Geometry of Complex Numbers. soe ee AS ‘De Moivre Formuta and the Calculation of Roots. 2 2... 48 ‘Complex Functions. Euler's Formula. toe ee Applications of Euler's Formula... | Multivalued Functions and Riemann Surfaces see ee SB Analytic Functions, Cauchy Theorem. . re Other Integral Theorems. Cauchy Integral Formula |. 1 |. 62 Complex Sequencesand Series 2. 1 6 2 ww ee. 6 Taylor and Laurent Series... 6 ee Zeros and Singularities . cee ee ee B The Residue Theorem and its Applications woe ee ee ee 8B Conformal Mapping by Analytic Functions. 2 2 2... OT Complex Sphere and PointatInfinity 2 2. 2... . (102 Integral Representations. 2... wee 108 Linear Differential Equations of Second Order General Introduction. The Wronskian . soe ee ee IDB General Solution of The Homogeneous Equation . woes 25 ‘The Nonhomogencous Equation. Variation of Constants. ||. 126 Power Series Solutions... 2... we ee OB The Frobenius Method. 2... we ee we 10 Some other Methods of Solution. 2 2 2 | 2. ws. OT viivi Chapter 4 4l 42 43 44 45 46 AT 48 Chapter 5 SA 3.2 5.3 5.4 55 5.6 57 5.8 59 5.10 Chapter 6 6.1 6.2 6.3 64 6.5 66 67 68 69 6.10 Chapter 7 a 72 73 TA 4S 76 W CONTENTS: Fourier Series Trigonometric Series 6. se Definition of Fourier Series 2 5 2 1 ee ee Examples of Fourier Series. . Loe Parity Properties, Sine and Cosine Series ‘Complex Form of Fourier Series . Le Pointwise Convergence of Fourier Series... Convergence in the Mean. Applications of Fourier Series . The Laplace Transformation Operational Calculus Co The Laplace Integral... Soe Basic Properties of Laplace Transform... 5s The Inversion Problem. . tee The Rational Fraction Decomposition : The Convolution Theorem. . Ls Additional Properties of Laplace Transform toe Periodic Functions. Rectification. 2. 6 6 ee ee The Mellin Inversion Integral... 6 1 2 ee Applications of Laplace Transforms . ‘Concepts of the Theory of Distributions Strongly Peaked Functions and The Dirac Delta Function Delia Sequences... co ‘The 8-Calculus : Representations of Delta Functions Applications of The 8-Calculus Weak Convergence . te Correspondence of Funetions and Distributions. Properties of Distributions . . Sequences and Series.of Distributions 5. 5. _ Distributions in N dimensions . : Fourier Transferms Representations of aFunction . 2 6 6 + - Examples of Fourier Transformations : Properties of Fourier Transforms... . ~ Fourier Integral Theorem . Lo Fourier Transforms of Distributions . 6. - + Fourier Sine and Cosine Transforms . . : Applications of Fourier Transforms. ‘The Prineiple “of Causality - 154 155 157 161 165 167 168 172, n9 180 184 187 189 194 200 206 210 224 223 226 229 232 236 240 245 250 287 260 262 266 269 27h 273 276 Chapter 8 8.1 B2 83 84 85 8.6 87 88 89 Chapter 9 91 92 93 94 95 9.6 OT 98 99 9.10 OAT Chapter 10 10.t 10.2 103 104 10.5 10.6 10.7 10.8 10.9 ° Chapter 11 wd 1.2 u3 114 11.5 6 “7 Ha CONTENTS, Partial Differential Equations The Stretched String. Wave Equation The Method of Separation of Variables . Laplace and Poisson Equations The Diffusion Equation . Use of Fourier and Laplace Transforms » The Method of Eigenfunction Expansions and Finite Transforms Continuous Eigenvalue Spectrum . Vibrations of a Membrane. Degeneracy . Propagation of Sound. Helmbeltz Equation Special Functions Cylindrical and Sphericat Coordinates The Common Boundary-Valuc Problems The Sturm-Liouville Problem . Self-Adjoint Operators Legendre Polynomials. Fourier-Legendre Series Bessel Functions : Associated Legendre Functions and Spherical Harmonies - Sphericat Bessel Functions . Neumann Functions Modified Bessel Functions - Finite-Dimensional Linear Spaces ‘Oscillations of Systems with Two Degrees of Freedom ‘Normal Coordinates and Linear Transformations Vector Spaces, Bases, Coordinates Lincar Operators, Matrices, Inverses . Changes of Basis Inner Product. Orthogonality. Unitary Operators - ‘The Metric. Generalized Orthogonality Eigenvalue Problems. Diagonalization Simultaneous Diagonalization . . . Jnfinite-Dimensional Vector Spaces Spaces of Functions : The Postulates of Quantum Mechanics ‘The Harmonic Oscillator Matrix Representations of Linear Operators Algchraic Methods of Solution Bases with Generalized Orthogonality Stretched String with a Discrete Mass in the Applications of Figenfunctions, 287 294 295 297 299 304 308 33 39 332 34 337 340 442 350 355 372 381 388 304 408 ait 419 424 433 437 441 443 451 463 467 47 476 483 488 492 495x CONTENTS Chapter 12 12.1 12.2 12.3 124 12.5 126 12.7 12.8 12.9 Chapter 13 ABA 13.2 133 13.4 13.5 13.6 13.7 13.8 13.9 Chapter 14 14.1 14.2 14.3 144 14.5 46 14.7 14.8 14.9 14.10 Chapter 15 151 15.2 15.3 15.4 15.5 15.6 ‘Chapter 16 16.1 16.2, ‘Green’s Functions Introduction... Lk Green's Function for the Sturm-Liouville Operator |. Series Expansions for Gx}... eee Green’s Functions in Two Dimensions Green's Functions for Initial Conditions . Green's Functions with Reflection Properties Green’s Functions for Boundary Conditions The Green’s Function Method 2 2. 2. ee A Case of Continuous Specttum © 2 2. eee Yariational Methods The Brachistochrone Problem . eee The Euler-Lagrange Equation. . 2... Hamilton's Principle... woe Problems involving Sturm-Liouvi Operators ve The Rayleigh-Ritz Method woe ee Variational Problems with Constraints. . Yariational Formulation of Eigenvatue Problems Variational Problems in Many Dimensions . . Formulation of Eigenvalue Problems by The Ratio Method ‘Traveling Waves, Radiation, Scattering Motion of Infinite Stretched Suing. - 2 Propagation of Initial Conditions - : Semi-infnite String, Use of Symmetry Properties. Energy and Power Flow in a Stretched String Generation of Waves in a Stretched String Radiation of Sound from a Pulsating Sphere : The Retarded Potential =. - woe Traveling Waves in Nonhomogeneous Media Scattering Amplitudes and Phase Shifts Seattering in Three Dimensions. Partial Wave Analysis Perturbation Methods ‘Introduction The Born Approximation . Le Perturbation of Eigenvalue Problems... First-Order Rayleigh-Schridinger Theory The Second-Order Nondegenerate Theory... - ‘The Case of Degenerate Eigenvalues... . . Tensors Introduction . . Two-Dimensional Stresses . 503 508 514 520 523 527 53 536 543 553 554 560 362 565 567 573 S17 581 589 592, 595 599 603 61! 619 624 628 633 644 647 650 653 658 665 ert 672 16.3 16.4 16.5 16.6 16.7 16.8 16.9 16.10 1601 Cartesian Tensors Algebra of Cartesian Tensors. . Kronecker and Levi-Civita Tensors. Psendotensors Derivatives of Tensors. Strain Tensor and Hooke’s Law . Tensors in Skew Cartesian Frames. Covariant and Contravariant Representations . General Tensors . Algebra of General ‘Tensors. Relative Tensore . The Covariant Derivative . Calculus of General Tensors Index . CONTENTS. xi 676 681 684 687 696 700 705 7 ns Rn?CHAPTER 1 VECTORS, MATRICES, AND COORDINATES 1.1 INTRODUCTION To be able to follow this text without undue difficulties, the reader is expected to have adequate preparation in mathematics and physics. ‘This involves a good working knowledge of advanced calculus, a basic course in differential equations, and a basic course in undergraduate algebra. Rudimentary knowledge of complex numbers, matrices, and Fourier series is very desirable but not indispensable. As for the subjects in physics, the reader should have completed the standard undergraduate. training in mechanics, thermodynamics, electromagnetism, and atomic physics. Despite these prerequisites, a need is often recognized for reviewing some of the preparatory material at the beginning of a text. Let us follow this custom and devote some time to the subject of vector analysis which has a bearing, in more than one way, on the material developed in this text. Of course, such a review must be brief and we must omit all the details, in particular those involving mathematical proofs. The reader is referred to standard textbooks in advanced calculus and vector analysis* for a full discussion. On the other hand, we hope to draw attention to some interesting puints not always emphasized in commonly used texts. 1.2 VECFORS IN CARTESIAN COORDINATE SYSTEMS In many elementary textbooks a vector is defined as a quantity characterized by magnitude and direction. We shalt see in Chapter 10 that vectors are much more general than this, but it is fair to say that the concept of vectors was first intro- duced into mathematics (by physicists) to represent “quantities with direction,” eg., displacement, velocity, force, ete. Doubtless, they are the simplest and most familiar kinds of vectors. As we well know, quantities with direction can be graphically represented by arrows and are subject to two basic operations: a) multiplication by a scalar,t —_ 6) addition. These operations are illustrated in Fig. 1.1. * For example, A. E. Taylor, Advanced Catculus; T. M. Apostol, Mathematical Analysis; W. Kaplan, Adeanced Calculus. + Until we are ready to discuss complex vectors (Chapler 10) we shall assume that scalars are reat numbers. 12. VECTORS, MATRICES, AND COORDINATES 12 we a“ “Oe In many cases we can plot various vectors from a single point, the origin. Then each vector can be characterized by the coordinates of its “tip.” Various coordinate systems are possible but the cartesian coordinate systems are the most convenient. The reason is very simple and very deep: The cartesian coordinates of a point can serve as the components of the corresponding vector at the same time. ‘This is illustrated in Fig. 1.2, where orthogonal cartesian systems, in plane and in space, are selected. Note that the three-dimensional system is “right-handed” ;* in general, we shall use right-handed systems in this book. Figure 1.1 Figure 1.2 Wecan now associate with a vector u (in space) a set of three scalars (ts, ty, ts), such that Au will correspond to (Autz, My, du.) and u-+ ¥ will correspond to (tz + Os, thy + dy, 4s + 0,). Note that no such relations hold, in general, if a vector is characterized by other types of coordinates, e.g., spherical or cylindrical. In addition, orthogonal cartesian coordinates result in very simple formulas for other common quantities associated with vectors, such as : a) length (ihagnitude) of a vector: Yo] = = ud tay + ay, b) projections of a vector on coordinate axes:f u, = ucos(u,i), ty = ucos (a,j), v4 = wos (u, k), * Rotation of the x-axis by 90° to coincide with the y-axis appears counterclodkwise for all observers with z > 0. — + Standard notation is used: The symbols, j, k are unit vectors in x-, y-, and z-directions, respectively. The symbot (u, ) stands for the angle between the directions given by wand ¥, 12 VECTORS IN CARTESIAN COORDINATE SYSTEMS 3 ©) projection of a vector on an arbitrary direc- = tion defined by vector s (Fig. 1.3): OP = nu, = ucosy = u, cos (s, i) + u, cos (5, j} + 1, cos (s, &), ¢ scalar product (dot product) of two vectors: (uy) = wos (a, ¥) = wb. + wy + Udy ¢) vector product (cross product): . fu X ¥] = Gp, — 420y)h + Qeds — 0,)j + (iedy — yP.)K. Figure 1.3 The important distinctive [eature of the cross product is that fu X v] ¥ [v X uh, namely, it is not commutative; rather, it is anticommutative: fu X vy] = -[fy X a} Remark. Apart from its important physical applications, the cross product of two vectors leads us to the concept of “oriented area.” The magnitude of [u x ¥), namely uo sin (u, »)], is equal lo the area of the parallelogram formed by wand ¥. The direction of [u X ¥] can serve to distinguish the “positive side” of the parallelogram from its “negative side.” Figure 1.4 shows two views of the same parallelogram illustrating this idea,” [ux] Positive side Figure 1.4 Closely related to this property is the concept of a right-handed triple of vectors. ,Any three vectors a, v, and w, taken in this order, are said to form a right- handed (or positive) triple if the so-called triple product (lu X vw) is positive.* This happens if w is on the same side of the plane defined by the vectors u and ¥, as illustrated in Fig. 1.5. It is not hard to verify that ((w X ¥] > w) represents, in this case, the volume V of the paralfetepiped formed by the vectors u, ¥, and wy, * These vectors form a left-handed (negalive) triple if (u X ¥]- w< 0.4 VECTORS, MATRICES, AND COORDINATES 1.3 Exercise. Show that = {la x vw] under any circumstances. Show also that the sign of the triple product is unchanged under cyclic permutation of u, ¥, and w, that is, (la Xv] -#) = (lr X why) = tly x wd). Figure 1.5 1.3 CHANGES OF AXES, ROTATION MATRICES We have seen that a given vector wis associated with a set of three numbers, namely its components,* with respect to some orthogonal cartesian system. However, it is clear that if the system of axes is changed, the components change as well. Let us study these changes. Consider, for vectors in a plane, a change in the system of axes which is pro- duced by a rotation by the angle 0, as illustrated in Fig. 1.6. The old system is (x, y) and the new system is (x, y’). Since u = tf + uyj, the x’-component of u is the sum of projections of vectors u,i and uj on the x’-axis, and similarly for the y’-component.{ From the diagram we see that this yields Jub = ucos@ + tysin®, = —uz sind 1 uycos 8. It is insteuctive to note that the angle between the x’- and y-axes is (w/2 — 4) while the angle between the y’- and x-axes is (x/2 + 6). In view of sind = cos (5 - ) ~sin 8 = cos (3 + ): we see that all four coefficients in the above equations represent cosines of the angles between the respective axes, Let us now turn to the three-dimensional case. Figure 1.7 represents two orthogonal cartesian systems, both right-handed, centered at O. It is intuitively clear that the primed system can be obtained from the unprimed one by the mo- tion of a “ body about a fixed point.” In fact, it is shown in almost any text- book on mechanics} that such a motion can be reduced to a rotation about some axis (Euler’s theorem). and ; * Instead of “components,” the term “coordinates of a vector” is often used (see also. Section 10,3). + For example, Goldstein, Classical Mechanics, Section 4.6. 1.3 CHANGES OF AXES. ROTATION MATRICES 5. Write u= «b+ uj + uk, collect contributions to ut from the three vectors wei, uyj, and ak, and obtain ue = uz cos (i, i) + 4, cos (i,j) -F ui, cos (V’, k), where ’ is, of course, the unit vector in the x’-direction. Note that the cosines involved are the directional cosines of the x‘-direction with respect to the unprimed system or, for that matter, the dot products of i’ with i, j, and k. It is clear that similar formulas can be written for uf and wf. At this stage, however, it is very convenient to switch to a different notation: Instead of writing (tgs Myr He), let Us write (144, wz, us) and similarly (14, «4, u4) for (ud, wh, 04). More- over, denote by an the angle between the mth primed axis and the nth unprimed axis (Aree such angles are marked on Fig. 1.6) and by dan the corresponding cosine (that is, Gan = COSam»). This new notation permits us to write the transformation formulas in an easily memorized pattern: Hy = 443th + Ayala + aygtg, us dais + aa2tt, + aagts, U3 = a3 ty + agatte + aagtts, or, if desired, in the compact form the = DO date Un = 3,2,3). mot From this analysis we conclude that the new components (i, 1, 4) can be obtained from the old components (i), #2, wa) with the help of nite coefficients.6 VECTORS, MATRICES, AND COORDINATES. 13 These nine coefficients, arranged in the self-explanatory pattern below are said to form a matrix.* We shall denote matrices by capital letters. Columns Ist 2nd 3rd Ist 2nd )Rows 3rd Matrix 4 has three rows and three columns; the individual coefficients dmx are referred to as matrix elements, or entries. It is customary to use the first subscript (mm in our case) to label the row and the second one to label the column; thus the matrix element a, should be located at the intersection of the kth row with the ith column. The set of elements in a given row is very often called a row vector and the set of elements in a column, a column vector. This nomenclature is justified by the fact that any three numbers can be treated as components of some vector in space. However, at this stage it is worthwhile to make a digression and establish a geo- metric interpretation for the column vectors of 4. The reader should pay par- ticular attention to the argument because of its general significance. . Let us imagine a unit vector u. Suppose the unprimed system was oriented in such a way that u was along the x-axis. Then the components of ware (1, 0, 0) and u actually coincides with the vector i. If the coordinate system is now rotated, the new components of u are given by uy = ah + a0 + 4190 = aus, Wy = Gah + 920 + dog) = G21, Wh = ag! + 2320 + 49:0 = as. We see that the first column vector of matrix 4 is composed of the new components of vector u. In other words, we can say that the new coniponenis of i are (@11, 421) 254) and we can write i= ayy + agi’ + ak’. Similar statements telate § and k to the second and third column vectors of 4. Note that in this discussion the unit vectors i, j,k assume a role independent of their respective coordinate axes, The axes are rotated but the vectors i, ik stay in place and are then referred to the rotated system of axes. * More precisely, a3 3 matrix is formed. The reader cant easily construct-an analogous 2% 2 matrix to account for two-dimensional rotations. 13 CHANGES OF AXES. ROTATION MATRICES 7 Exercise. Establish the geometrical meaning of the row vectors of matrix 4 representing a rotation. The definitions introduced above allow the computation of (u4, u>, 43) from (2, ta, 3) by the following rule: To obtain uj, take the dot product of the kth row of matrix A with the vector u, as given by the triple (u,, 2, u3). Since in this process each row of the matrix is “dotted” with (ws, #2, 43), we may regard it as some kind of multiplication of a vector by a matrix. In fact, this operation is commonly known as vector-matrix multiplication and is visually exhibited as shown: um wh ip = ty us us Matrix A Column vector # Column vector u’ As we see, the old components are arranged in a column which we shall denote by u. This column vector is multiplied by the matrix A and this results in another column vector, denoted by u’. The multiplication means, of course: Form the dot product of the first row of A with u for the first component of u'; then form the dot product of the second row of A with u to get uh, and similarly for uy. The entire procedure is symbolically written as Awa. Remark, Note that the set (w), #2, #3), arranged in a column, has not been denoted simply by u but rather by 2 new symbol w.* The point is that in the context of our problem, both w and u represent the same vector u, but with respect to different systems of axes. We must think of u and u’ as two different representations of u, and the matrix A shows us how to switch from one representation to another, Before we discuss further topics involving matrices, let us record the fact that our matrix A is not just a collection of nine arbitrary scalars. Its matrix ele- ments are interdependent and possess the following properties. a) The columns of A are orthogonal to each other, namely, 1112 F a2idag + 51932 = 0, 442019 + ae2t23 + ay2435 = 0, 419411 + Aoadas + G3a9g1 = 0. This property follows from the fact that the colunins of A are representations (in the new system) of the vectors i, j, and k and these vectors are mutually orthogonal. * The symbol » should not be confused with {ul, the magnitude of vector u, which is also denoted by (p. 2).