Tensor Analysis and Curvilinear Coordinates

Phil Lucht
Rimrock Digital Technology, Salt Lake City, Utah 84103
last update: May 19, 2016

Maple code is available upon request. Comments and errata are welcome.
The material in this document is copyrighted by the author.

The graphics look ratty in Windows Adobe PDF viewers when not scaled up, but look just fine in this
excellent freeware viewer: http://www.tracker-software.com/pdf-xchange-products-comparison-chart .

The table of contents has live links. Most PDF viewers provide these links as bookmarks on the left.

Overview and Summary ......................................................................................................................... 7

1. The Transformation F: invertibility, coordinate lines, and level surfaces.................................. 12
Example 1: Polar coordinates (N=2).................................................................................................. 13
Example 2: Spherical coordinates (N=3) ........................................................................................... 14
Cartesian Space and Quasi-Cartesian Space ....................................................................................... 15
Pictures A,B,C and D.......................................................................................................................... 16
Coordinate Lines................................................................................................................................. 16
Example 1: Polar coordinates, coordinate lines ................................................................................. 17
Example 2: Spherical coordinates, coordinate lines .......................................................................... 18
Level Surfaces..................................................................................................................................... 18
2. Linear Local Transformations associated with F : scalars and two kinds of vectors ................. 20
2.1 Linear Local Transformations....................................................................................................... 20
2.2 Scalars ........................................................................................................................................... 21
2.3 Contravariant vectors .................................................................................................................... 22
2.4 Covariant vectors .......................................................................................................................... 22
2.5 Bar notation................................................................................................................................... 23
2.6 Origin of the names contravariant and covariant .......................................................................... 23
2.7 Other vector types? ....................................................................................................................... 24
2.8 Linear transformations .................................................................................................................. 25
2.9 Vectors that are contravariant by definition.................................................................................. 26
2.10 Vector Fields............................................................................................................................... 26
2.11 Names and symbols .................................................................................................................... 27
2.12 Definition of the words "scalar", "vector" and "tensor".............................................................. 28
3. Tangent Base Vectors en and Inverse Tangent Base Vectors u'n .................................................. 30
3.1 Differential Displacements ........................................................................................................... 30
3.2 Definition of the en ; the en are the columns of S......................................................................... 31
3.3 en as a contravariant vector ........................................................................................................... 33
3.4 A semantic question: unit vectors ................................................................................................ 33
Example 1: Polar coordinates, tangent base vectors .......................................................................... 34
Example 2: Spherical Coordinates, tangent base vectors................................................................... 35
3.5 The inverse tangent base vectors u'n and inverse coordinate lines................................................ 36
Example 1: Polar coordinates: inverse tangent base vectors and inverse coordinate lines............... 37
1
4. Notions of length, distance and scalar product in Cartesian Space.............................................. 39
5. The Metric Tensor ............................................................................................................................ 41
5.1 The Picture D Context .................................................................................................................. 41
5.2 Definition of the metric tensor ...................................................................................................... 41
5.3 Inverse of the metric tensor........................................................................................................... 44
5.4 A metric tensor is symmetric ........................................................................................................ 44
5.5 det(g) and gnn of a Cartesian-generated metric tensor are non-negative....................................... 45
5.6 Definition of two kinds of rank-2 tensors ..................................................................................... 45
5.7 Proof that the metric tensor and its inverse are both rank-2 tensors ............................................. 46
5.8 Metric tensor converts vector types .............................................................................................. 49
5.9 Vectors in Cartesian space ............................................................................................................ 49
5.10 The covariant dot product A • B and norm |A| ......................................................................... 50
5.11 Metric tensor and tangent base vectors: scale factors and orthogonal coordinates................... 52
5.12 The Jacobian J............................................................................................................................. 55
5.13 Some relations between g, R and S in Pictures B and C (Cartesian x-space). ............................ 58
Example 1: Polar coordinates: metric tensor and Jacobian............................................................... 60
Example 2: Spherical coordinates: metric tensor and Jacobian ........................................................ 61
5.14 Special Relativity and its Metric Tensor: vectors and spinors ................................................... 62
5.15 General Relativity and its Metric Tensor .................................................................................... 65
5.16 Continuum Mechanics and its Metric Tensors............................................................................ 66
6. Reciprocal Base Vectors En and Inverse Reciprocal Base Vectors U'n........................................ 73
6.1 Definition of the En ....................................................................................................................... 73
6.2 The en and En Dot Products and Reciprocity (Duality) ................................................................ 74
6.3 Covariant partners for en and En .................................................................................................. 78
6.4 Summary of the basic facts about en and En ................................................................................. 80
6.5 Repeat the above for the inverse transformation: definition of the U'n ........................................ 80
6.6 Expanding vectors on different sets of basis vectors .................................................................... 81
6.7 Another way to write the En .......................................................................................................... 84
6.8 Comparison of ē n and En .............................................................................................................. 85
6.9 Handedness of coordinate systems: the en , the sign of det(S), and Parity .................................. 86
7. Translation to the Standard Notation ............................................................................................. 89
7.1 Outer Products .............................................................................................................................. 89
7.2 Mixed Tensors and Notation Issues .............................................................................................. 90
7.3 The up/down bell goes off ............................................................................................................ 91
7.4 Some Preliminary Translations; raising and lowering tensor indices with g ............................... 92
7.5 Dealing with the matrices R and S ; various Rules and Theorems ............................................... 96
7.6 Orthogonality Rules, Inversion Rules, Cancellation Rules......................................................... 101
7.7 About δ and ε .............................................................................................................................. 104
7.8 Covariance and Matrix Multiplication ........................................................................................ 105
7.9 Matrix Inverse, Transpose and Determinant............................................................................... 108
7.10 Tensors of Rank n, direct products, Lie groups, symmetry and Ricci-Levi-Civita .................. 114
7.11 The Contraction Tilt-Reversal Rule .......................................................................................... 117
7.12 The Contraction Neutralization Rule ........................................................................................ 119
7.13 The tangent and reciprocal base vectors and expansions on same............................................ 121
7.14 Comment on Covariant versus Contravariant ........................................................................... 125
7.15 The Significance of Tensor Analysis ........................................................................................ 126
2
 7.16 The Christoffel Business: covariant derivatives ...................................................................... 129
7.17 Expansions of higher order tensors ........................................................................................... 130
7.18 Collection of Facts about basis vectors en , u'n and bn. ............................................................. 131
7.19 More on basis vectors and matrix elements of R and S ............................................................ 134
8. Transformation of Differential Length, Area and Volume ......................................................... 142
8.1 Overview of Chapter 8................................................................................................................ 142
8.2 The differential N-piped mapping............................................................................................... 143
8.3 Properties of the finite N-piped spanned by the en in x-space .................................................... 145
8.4 Back to the differential N-piped mapping: how edges, areas and volume transform ................ 147
(a) The Setup................................................................................................................................. 147
(b) Edge Transformation............................................................................................................... 148
(c) Area Transformation................................................................................................................ 149
(d) Volume Transformation .......................................................................................................... 150
(e) Covariant Magnitudes.............................................................................................................. 151
(f) Two Theorems : g'nn g' = cof(g'nn) and |(Πxi≠nei)| = cof(g'nn) ...................................... 152
(g) Cartesian-View Magnitude Ratios .......................................................................................... 154
(h) Nested Cofactor Formulas and STS notation ........................................................................... 155
(i) Transformation of arbitrary differential vectors, areas and volume......................................... 158
(j) Concatenation (Composition) of Transformations................................................................... 160
(k) Examples of area magnitude transformation for N = 2,3,4 ..................................................... 161
Example 2: Spherical Coordinates: area patches ........................................................................ 162
8.5 Transformation of Differential Volume applied to Integration................................................... 163
8.6 Interpretations of the Jacobian .................................................................................................... 165
8.7 Volume integration of a tensor field under linear transformations ............................................. 165
9. The Divergence in curvilinear coordinates ................................................................................... 167
9.1 Geometric Derivation of the Curvilinear Divergence Formula .................................................. 167
9.2 Various expressions for div B..................................................................................................... 170
9.3 Translation from Picture B to Picture M&S ............................................................................... 172
9.4 Comparison of various authors' notations................................................................................... 173
10. The Gradient in curvilinear coordinates..................................................................................... 175
10.1 Expressions for grad f ............................................................................................................... 175
10.2 Expressions for grad f • B......................................................................................................... 178
11. The Laplacian in curvilinear coordinates ................................................................................... 180
12. The Curl in curvilinear coordinates ............................................................................................ 182
12.1 Definition of curl B................................................................................................................... 182
12.2 Computation of the line integral ............................................................................................... 183
12.3 Solving for the curl ................................................................................................................... 185
12.4 Various forms of the curl .......................................................................................................... 186
12.5 The curl in orthogonal coordinate systems ............................................................................... 188
12.6 The curl in N > 3 dimensions.................................................................................................... 189
13. The Vector Laplacian in curvilinear coordinates....................................................................... 192
13.1 Derivation of the Vector Laplacian in general curvilinear coordinates .................................... 192
13.2 The Vector Laplacian in orthogonal curvilinear coordinates.................................................... 195
13.3 The Vector Laplacian in Cartesian coordinates ........................................................................ 196

3
14. Summary of Differential Operators in curvilinear coordinates ............................................... 198
14.1 Summary of Conventions and How To..................................................................................... 198
14.2 divergence ................................................................................................................................. 200
14.3 gradient and gradient dot vector................................................................................................ 200
14.4 Laplacian................................................................................................................................... 201
14.5 curl ............................................................................................................................................ 202
14.6 vector Laplacian........................................................................................................................ 203
14.7 Example 1: Polar coordinates: a practical curvilinear notation.............................................. 204
15. Covariant derivation of all curvilinear differential operator expressions ............................... 206
15.1 Review of Chapters 9 through 13 ............................................................................................. 206
15.2 The Covariant Method .............................................................................................................. 207
15.3 divergence (Chapter 9).............................................................................................................. 209
15.4 gradient and gradient dot vector (Chapter 10) .......................................................................... 209
15.5 Laplacian (Chapter 11).............................................................................................................. 210
15.6 curl (Chapter 12) ....................................................................................................................... 211
15.7 vector Laplacian (Chapter 13)................................................................................................... 211
15.8 Verification that two tensorizations are the same .................................................................... 214
Appendix A: Reciprocal Base Vectors the Hard Way .................................................................... 217
A.1 Introduction................................................................................................................................ 217
A.2 Definition of En .......................................................................................................................... 218
A.3 Simpler notation......................................................................................................................... 218
A.4 Generalized Cross Product of N-1 vectors of dimension N ....................................................... 218
A.5 Missing Man Formation............................................................................................................. 220
A.6 Apply this Notation to E ............................................................................................................ 220
A.7 Compute Em • en ......................................................................................................................... 221
A.8 Compute En • Em ........................................................................................................................ 222
A.9 Summary of relationship between the tangent and reciprocal base vectors............................... 222
A.10 Another Cross Product Notation and another expression for E ............................................... 223
Appendix B: The Geometry of Parallelepipeds in N dimensions................................................... 224
B.1 Overview .................................................................................................................................... 224
B.2 Preliminary: Equation of a plane in N dimensions.................................................................... 224
B.3 N-pipeds and their Faces in Various Dimensions ...................................................................... 225
(a) The 1-piped.............................................................................................................................. 225
(b) The 2-piped ............................................................................................................................. 225
(c) The 3-piped.............................................................................................................................. 227
(d) The N-piped............................................................................................................................. 228
B.4 The question of inward versus outward facing normal vectors.................................................. 229
B.5 The Face Area and Volume of N-pipeds in Various Dimensions ............................................. 230
(a) The 2-piped.............................................................................................................................. 230
(b) The 3-piped ............................................................................................................................. 231
(c) The 4-piped.............................................................................................................................. 234
(d) The N-piped............................................................................................................................. 235
B.6 Summary of Main Results of this Appendix .............................................................................. 237
Appendix C: Elliptical Polar Coords, Views of x'-space, Jacobian Integration Rule................. 239
C.1 Elliptical polar coordinates......................................................................................................... 239
C.2 Forward coordinate lines ............................................................................................................ 240

4
 C.3 Inverse coordinate lines.............................................................................................................. 240
C.4 Drawing a contravariant vector V in x-space: the meaning of V'n ........................................... 241
C.5 Drawing a contravariant vector V' in x'-space: two "Views" .................................................... 242
C.6 Drawing the specific contravariant vector dx in x-space and x'-space....................................... 245
C.7 Study of how dx transforms in the mapping between x-space and x'-space .............................. 247
C.8 A Derivation of the Jacobian Integration Rule........................................................................... 249
Appendix D: Tensor Densities and the ε tensor .............................................................................. 252
D.1 Definition of a tensor density..................................................................................................... 252
D.2 A few facts about tensor densities.............................................................................................. 253
D.3 Theorem about Totally Antisymmetric Tensors: there is really only one: εabc... ................... 256
D.4 The contravariant ε tensor .......................................................................................................... 257
D.5 Some facts about the ε tensor..................................................................................................... 259
D.6 The covariant ε tensor : repeat Section D.4 as if its weight were not known ........................... 261
D.7 Generalized cross products ........................................................................................................ 262
D.8 The tensorial nature of curl B..................................................................................................... 262
D.9 Tensor E as a weight 0 version of ε : three conventions ........................................................... 263
D.10 Representation of ε, εε and contracted εε as determinants...................................................... 267
D.11 Covariant forms of the previous Section results ...................................................................... 274
D.12 How determinants of rank-2 tensors transform........................................................................ 277
Appendix E: Tensor Expansions: direct product, polyadic and operator notation.................... 281
E.1 Direct Product Notation.............................................................................................................. 281
E.2 Tensor Expansions and Bases..................................................................................................... 282
E.3 Polyadic Notation ....................................................................................................................... 288
E.4 Dyadic Products.......................................................................................................................... 289
E.5 Matrix notation for dyadics (Cartesian Space) ........................................................................... 290
E.6 Large and small dots used with dyadics (Cartesian Space) ........................................................ 291
E.7 Operators and Matrices for Rank-2 tensors: the bra-ket notation (Cartesian Space) ................ 292
E.8 Expansions of tensors on unit tangent base vectors: M and N .................................................. 298
E.9 Application of Section E.8 to Orthogonal Curvilinear Coordinates........................................... 303
E.10 Tensor expansions in a mixed basis ......................................................................................... 309
Appendix F: The Affine Connection Γcab and Covariant Derivatives .......................................... 311
F.1 Definition and Interpretation of Γ : Γcab = qc • (∂aqb) = Rci(∂aRbi)..................................... 311
F.2 Identities of the form (∂aRdn) = – Ren Rdm (∂aRem).................................................................... 313
F.3 Identities of the form (∂cgab) = – [gan Γ bcn + gbn Γacn] ......................................................... 314
F.4 Identity: Γdab = (1/2) gdc [ ∂agbc + ∂bgca – ∂cgab] ................................................................. 316
F.5 Picture D1 Context ..................................................................................................................... 319
F.6 Relations between Γ and Γ '........................................................................................................ 320
F.7 Statement and Proof of the Covariant Derivative Theorem........................................................ 320
F.8 Rules for raising any index on a covariant derivative of a covariant tensor density. ................. 326
F.9 Examples of covariant derivative expressions............................................................................ 328
F.10 The Leibniz rule for the covariant derivative of the product of two tensor densities ............... 331
Appendix G: Expansion of (∇v) in curvilinear coordinates (v = vector) ...................................... 335
G.1 Continuum Mechanics motivation ............................................................................................. 335
G.2 Expansion of ∇v on ei⊗ej by Method 1: Use the fact that vb;a is a tensor.............................. 336
G.3 Expansion of ∇v on ei⊗ej by Method 2: Use brute force. ....................................................... 338

5
 G.4 Expansion on ei⊗ej and ^ei⊗e^j.................................................................................................. 339
G.5 Orthogonal coordinate systems .................................................................................................. 340
G.6 Maple evaluation of (∇v) in several coordinate systems ........................................................... 341
Appendix H: Expansion of div(T) in curvilinear coordinates (T = rank-2 tensor) ...................... 344
H.1 Introduction................................................................................................................................ 344
H.2 Continuum Mechanics motivation ............................................................................................. 344
H.3 Expansion of divT on en by Method 1: Use fact that Tab;α is a tensor. .................................... 345
H.4 Expansion of divT on en by Method 2: Use brute force............................................................ 346
H.5 Adjustment for T expanded on (e ^i⊗e ^j) and divT expanded on ^ e a ........................................... 348
H.6 Maple: divT in cylindrical and spherical coordinates ............................................................... 349
Appendix I : The Vector Laplacian in Spherical and Cylindrical Coordinates........................... 351
I.1 Introduction ................................................................................................................................. 351
I.2 Method 1 : a review .................................................................................................................... 352
I.3 Method 1 for spherical coordinates: Maple speaks .................................................................... 354
I.4 Method 1 for spherical coordinates: putting results in traditional form ..................................... 357
I.5 Method 2, Part A ......................................................................................................................... 359
I.6 Method 2, Part B.......................................................................................................................... 360
I.7 Method 2 for spherical coordinates: Maple speaks again........................................................... 362
I.8 Results for Cylindrical Coordinates from both methods ............................................................. 364
Appendix J: Expansion of (∇T) in curvilinear coordinates (T = rank-2 tensor).......................... 368
J.1 Total time derivative as prototype equation ................................................................................ 368
J.2 Computation of components (∇T)'ijk ......................................................................................... 369
J.3 Tensor expansions of ∇T on the un and en base vectors.............................................................. 370
J.4 Tensor expansions of ∇T on the ^en base vectors......................................................................... 371
J.5 Total time derivative equation written in unit-base-vector curvilinear components ................... 372
J.6 Shorthand notations and a continuum mechanics application..................................................... 374
J.7 Maple computation of the (∇T)'ijk components for spherical coordinates................................. 376
J.8 Maple computation of the (∇T)'ijk components for cylindrical coordinates.............................. 380
J.9 The Lai Method of computing (∇T)'ijk for orthogonal coordinates........................................... 381
Appendix K: Deformation Tensors in Continuum Mechanics ...................................................... 385
K.1 A Preliminary Deformation Flow Picture .................................................................................. 385
K.2 A More Complicated Deformation Flow Picture....................................................................... 391
K.3 Covariant form of a solid constitutive equation involving the deformation tensor.................... 396
K.4 Some fluid constitutive equations .............................................................................................. 397
K.5 Corotational and other objective time derivatives of the Cauchy stress tensor.......................... 398
References ............................................................................................................................................ 406

6
 Overview

Overview and Summary

In this lengthy monograph, tensor analysis (also known as tensor algebra or tensor calculus) is developed
starting from Square Zero which is an arbitrary invertible continuous transformation x' = F(x) in N
dimensions.
The subject was "exposed" by Gregorio Ricci in the late 1800's under the name "absolute differential
calculus". He and his student Tullio Levi-Civita published a masterwork on the subject in 1900 (see
References). Christoffel and others had laid the groundwork a few decades earlier.
Three somewhat different applications of tensor analysis are treated concurrently.
Our primary concern is the subject of curvilinear coordinates in N dimensions. All the basic
expressions for the standard differential operators in general curvilinear coordinates are derived from
scratch (in several ways). These results are often stated but not so often derived.
The second application involves transformations connecting "frames of reference". These
transformations could be spatial rotations, Galilean transformations, the Lorentz transformations of
special relativity, or the transformations involving the effects of gravity in general relativity. Beyond
establishing the tensor analysis formalism, not much is said about this set of applications.
The third application deals with material flows in continuum mechanics.
The first six Chapters develop the theory of tensor analysis in a simple developmental notation where
all indices are subscripts, just as in normal college physics. After providing motivation, Chapter 7
translates this developmental notation into the Standard Notation in use today. Chapter 8 treats
transformations of length, area and volume and then the curvilinear differential operator expressions are
derived, one per Chapter, with a summary in the penultimate Chapter 14. The final Chapter 15 rederives
all the same results using the notion of covariance and associated covariant derivatives.
The information is presented informally as if it were a set of lectures. Little attention is paid to strict
mathematical rigor. There is no attempt to be concise: examples are given, tangential remarks are
inserted, almost all claims are derived in line, and there is a certain amount of repetition. The material is
presented in a planned sequence to minimize the need for forward references, but the sequence is not
perfect. The interlocking pieces of tensor analysis do seem to exhibit a certain logical circularity.
The reader will find in this document very many worked-out, detailed calculations, derivations and
proofs. The intent is to provide the reader with "hands-on experience" in working with all the tensor tools.
The only real prerequisites for the reader are a knowledge of calculus of several variables (such as the
chain rule and meaning of ∇2) and of basic linear algebra (matrices and determinants).

7
 Overview

The following Table of Contents (TOC) highlights in bold the main topics which are harder to see in the
detailed TOC presented above. In most PDF viewers, the clickable TOC entries appear also as clickable
bookmarks to the left of the text. Our equation references are not clickable, but the constant drumbeat of
sequential equation numbers on the right should make them easy to locate. (Split-screen PDF viewing
can be simulated by loading up two copies of the same PDF under different names, a very useful trick.)

1. The Transformation F: invertibility, coordinate lines, and level surfaces

2. Linear Local Transformations associated with F : scalars and two kinds of vectors
3. Tangent Base Vectors en and Inverse Tangent Base Vectors u'n
4. Notions of length, distance and scalar product in Cartesian Space
5. The Metric Tensor
6. Reciprocal Base Vectors En and Inverse Reciprocal Base Vectors U'n
7. Translation to the Standard Notation
8. Transformation of Differential Length, Area and Volume
9. The Divergence in curvilinear coordinates
10. The Gradient in curvilinear coordinates
11. The Laplacian in curvilinear coordinates
12. The Curl in curvilinear coordinates
13. The Vector Laplacian in curvilinear coordinates
14. Summary of Differential Operators in curvilinear coordinates
15. Covariant derivations of all curvilinear differential operator expressions
Appendix A: Reciprocal Base Vectors the Hard Way
Appendix B: The Geometry of Parallelepipeds in N dimensions
Appendix C: Elliptical Polar Coords, Views of x'-space, Jacobian Integration Rule
Appendix D: Tensor Densities and the ε tensor
Appendix E: Tensor Expansions: direct product, polyadic and operator notation
Appendix F: The Affine Connection Γcab and Covariant Derivatives
Appendix G: Expansion of (∇v) in curvilinear coordinates (v = vector)
Appendix H: Expansion of div(T) in curvilinear coordinates (T = rank-2 tensor)
Appendix I : The Vector Laplacian in Spherical and Cylindrical Coordinates
Appendix J: Expansion of (∇T) in curvilinear coordinates (T = rank-2 tensor)
Appendix K: Deformation Tensors in Continuum Mechanics
References

Here then is a brief summary of each Chapter and each Appendix.

Chapter 1 introduces the notion of the general invertible transformation x' = F(x) as a mapping
between x-space and x'-space. The range and domain of this mapping are considered in the familiar
examples of polar and spherical coordinates. These same examples are used to illustrate the general ideas
of coordinate lines and level surfaces. Certain Pictures are introduced to allow different names for the two
inter-mapped spaces, for the function F, and for its associated objects.
Chapter 2 introduces the linear transformations R and S=R-1 which approximate the (generally non-
linear) x' = F(x) in the local neighborhood of a point x. It is shown that two types of vectors naturally
arise in the context of this linearization, called contravariant and covariant, and an overbar is used to

8
 Overview

distinguish a covariant vector. Vector fields are defined and their transformations stated. The idea of
scalars and vectors as tensors of rank 0 and rank 1 is presented.
Chapter 3 defines the tangent base vectors en(x) which are tangent to the x'-coordinate lines in x-
space. In the example of polar coordinates it is shown that e = ^
r r and e = r ^
θ θ . The vectors e exist in x-
n
space and form there a complete basis which in general is non-orthogonal. The tangent base vectors
u'n(x') of the inverse transformation x = F-1(x') are also defined.
Chapter 4 gives a brief review of the notions of norm, metric and scalar product in Cartesian Space.
Chapter 5 addresses the metric tensor, called ḡ in x-space and ḡ' in x'-space. The metric tensor is
first defined as a matrix object ḡ, and then g ≡ ḡ -1. A definition is given for two kinds of (pure) rank-2
tensors (both matrices), and it is then shown that ḡ transforms as a covariant rank-2 tensor while g is a
contravariant rank-2 tensor. It is demonstrated how ḡ applied to a contravariant vector V produces a
vector that is covariant V̄= ḡ V, and conversely g V̄ = V. In Cartesian space g = 1, so the two types of
vectors coincide. The role of the metric tensor in the covariant vector dot product is stated, and the metric
tensor is related to the tangent base vectors of Chapter 3. The Jacobian J and associated functions are
defined, though the significance of J is deferred to Chapter 8. The last Sections briefly discuss the
connection between tensor algebra and special relativity (with a mention of spinor algebra), general
relativity, and continuum mechanics.
Chapter 6 introduces the reciprocal (dual) base vectors En which are later called en in the Standard
Notation. Of special interest are the covariant dot products among the en and En. It is shown how an
arbitrary vector can be expanded onto different basis sets. It is found that when a contravariant vector in
x-space is expanded on the tangent base vectors en, the vector components in that expansion are in fact
those of the contravariant vector in x'-space, V'i = RijVj. This fact proves useful in later Chapters which
express differential operators in x-space in terms of curvilinear coordinates and objects of x'-space. The
reciprocal base vectors U'n of the inverse transformation are also discussed.
Chapter 7 motivates and then makes the transition from the developmental notation to the Standard
Notation where contravariant indices are up and covariant ones are down. Although such a transition
might seem completely trivial, many confusing issues do arise. Once a matrix can have up and down
indices, matrix multiplication and other matrix operations become hazy: a matrix becomes four different
matrices. The matrices R and S act like tensors, but are not tensors, and in fact are not even located in a
well-defined space. Section 7.15 discusses the significance of tensor analysis with respect to physics in
terms of covariant equations, and Section 7.16 broaches the topic of the covariant derivative of a vector
field with its associated Christoffel symbols. Section 7.17 describes how to expand tensors of any rank in
various bases and notations, and finally Section 7.18 summarizes information about the basis vectors.
The focus then fully shifts to curvilinear coordinates as an application of tensor analysis. The final
Chapters are all written in the Standard Notation.
Chapter 8 shows how differential length, area and volume transform under x' = F(x). It considers the
inverse mapping of a differential orthogonal N-piped (N dimensional parallelepiped) in x'-space to a
skewed one in x-space. It is shown how the scale factors h'n = g'nn describe that ratio of N-piped edges,
while the Jacobian J = det(g'nn) describes the ratio of N-piped volumes. The relationship between the
vector areas of the N-pipeds is more complicated, and it is found that the ratio of vector area magnitudes
is cof(g'nn) . Heavy use is made of the results of Appendices A and B, as outlined below.
Chapters 9 through 13 use the information of Chapter 8 and earlier material to derive expressions
for all the standard differential operators expressed in general non-orthogonal curvilinear coordinates:
divergence, gradient, Laplacian, curl, and vector Laplacian. The last two operators are treated only in N=3

9
 Overview

dimensions where the curl has a vector representation, but then the curl is generalized to N dimensions.
For each differential operator, simplified results for orthogonal coordinates are also stated.
Chapter 14 gathers all the differential operator expressions into a set of tables, and revisits the polar
coordinates example one last time to illustrate a reasonably clean and practical curvilinear notation.
Chapter 15 rederives the general results of Chapters 9 through 13 using the ideas of covariance and
covariant differentiation. These derivations are elegantly brief, but lean heavily on the idea of tensor
densities (Appendix D) and on the implications of covariance of tensor objects involving covariant
derivatives (Appendix F).

About half of our content is contained in a set of Appendices.

Appendix A develops an alternative expression for the reciprocal base vector En as a generalized
cross product of the tangent base vectors en, applicable when x-space is Cartesian. This alternate En is
shown to match the En defined in Chapter 6, and the covariant dot products involving En and en are
verified.
Appendix B presents the geometry of a parallepiped in N dimensions (called an N-piped). Using the
alternate expression for En developed in Appendix A, it is shown that the vector area of the nth pair of
faces on an N-piped spanned by the en is given by ± An, where An = |det(S)| En , revealing a geometric
significance of the reciprocal base vectors. Scaled by differentials so dAn = |det(S)| En(Πi≠n dx'i), this
equation is then used in Chapter 9 where the divergence of a vector field is defined as the total flux of that
field flowing out through all the faces of the skewed differential N-piped in x-space divided by its
volume. This same dAn appears in Chapter 8 with regard to the transformation of N-piped face vector
areas.
Appendix C presents a case study of an N=2 non-orthogonal coordinate system, elliptical polar
coordinates. Both the forward and inverse coordinate lines are displayed. The meaning of the curvilinear
(x'-space) component V'n of a contravariant vector is explored in the context of this system, and the
difficulties of drawing such components in non-Cartesian (curvilinear) x'-space are pondered. Finally, the
Jacobian Integration Rule for changing integration variables is derived.
Appendix D discusses tensor densities and their rules of the road. Special attention is given to the
Levi-Civita ε tensor, including a derivation of all the εε contraction formulas and their covariant
statements. It is noted that the curl of a vector is a vector density.
Appendix E describes direct product and polyadic notations (including dyadics) and shows how to
expand tensors (and tensor densities) of arbitrary rank on an arbitrary basis.
Appendix F deals with covariant derivatives and the affine connection Γ which tells how the tangent
base vectors en(x') change as x' changes. Everything is derived from scratch and the results provide the
horsepower to make Chapter 15 go.
The next four appendices provide demonstrations of most ideas presented in this paper. In each
Appendix, a connection is made to continuum mechanics, and the results are then derived by "brute
force", by the covariant technique enabled by Appendix F, or by both methods. These Appendices were
motivated by the continuum mechanics text of Lai, Rubin and Krempl (referred to as "Lai", see
References). In each Appendix, it is shown how to express the object of interest in arbitrary curvilinear
coordinates. Maple code is provided for the general calculations, and that code is then checked to make
sure it accurately replicates the results quoted in Lai for spherical and cylindrical coordinates.
Appendix G treats the dyadic object (∇v), where v is a vector.
Appendix H does the same for the vector object divT where T is a rank-2 tensor.

10
 Overview

Appendix I deals with the vector Laplacian @B where B is a vector.

Appendix J treats the object (∇T), where T is a rank-2 tensor.
Appendix K (following Lai) discusses deformation tensors used in continuum mechanics. Those that
are truly tensors may be used to construct covariant constitutive equations describing continuous
o Δ ∇
materials. The covariant time derivatives Tn, Tn and Tn are derived ( n = 1 are objective Cauchy stress
rates). Some examples of covariant constitutive equations involving these tensor objects are listed.

Notations
RHS, LHS refer to the right hand side and left hand side of an equation
QED = which was to be demonstrated ("thus it has been proved")
a ≡ stuff means a is defined by stuff.
// indicates a comment on something shown to the left of //
diag(a,b,c..) means a diagonal matrix with diagonal elements a,b,c..
det(A), AT = determinant of the matrix A, transpose of a matrix A
Maple = a computer algebra system similar to Mathematica and MATLAB
vectors (like v) are indicated in bold; all other tensors (like T) are unbolded.
^
n = un = unit vector pointing along the nth positive axis of some coordinate system
∂i = ∂/∂xi and ∂t = ∂/∂t are partial derivatives
dt ≡ Dt ≡ D/Dt ≡ d/dt is a total time derivative, as in dtf(x,t) = ∂tf(x,t) + [∂f(x,t)/∂xi][∂xi/∂t]
V,a means ∂aV which means ∂V/∂xa, and V;a refers to the corresponding covariant derivative
x • y is a covariant dot product = gabxayb, except where otherwise indicated
when an equation is repeated after its initial appearance, its equation number is shown in italics
Figures are treated as if they were equations in terms of equation numbering
Components of a vector are written Vi = (V)i (bold used only in the second form)
Chapter N consists of Sections N.1, N.2 .....

11
 Chapter 1: The Transformation F

1. The Transformation F: invertibility, coordinate lines, and level surfaces

If x and x' are elements of the vector space RN (N-dimensional reals), one can specify a mapping

x' = F(x) F: RN → RN (1.1)

defined by a set of N continuous (C2) real functions Fi , each of N real variables,

x'1 = F1(x1, x2, x3... xN)

x'2 = F2(x1, x2, x3... xN)
...
x'N = FN(x1, x2, x3... xN) . (1.2)

If all functions Fi are linear in all of their arguments, then the mapping F: RN → RN is a linear mapping.
Otherwise the mapping is non-linear.
A mapping is often referred to as a transformation. We shall be interested only in transformations
which are 1-to-1 and are therefore invertible. For such transformations,

x' = F(x) x = F-1(x') or

x' = x'(x) x = x(x') . (1.3)

In the transformation x' = F(x), if x roams over the entire RN of x-space (the domain is RN), we may find
that x' roams over only some subset of RN in x'-space. The 1-to-1 invertible mapping is then between the
domain of mapping F which is all of RN, and the range of mapping F which is this subset.
As just noted, it will be assumed that x' = F(x) is essentially invertible so x = F-1(x') exists for any x'.
By essentially is meant there may be a few problem points in the transformation which can be "fixed up"
in some reasonable manner so that x' = F(x) is invertible (see examples below).
The functions Fi must be C1 continuous to support the linearization derivatives appearing in Chapter
2, and they must be C2 continuous to support some of the differential operators expressed in curvilinear
coordinates in Chapters 9-14.

12
 Chapter 1: The Transformation F

Example 1: Polar coordinates (N=2)

(a) The transformation from Cartesian to polar coordinates is given by,

x = (x1, x2 ) = (x,y)
x' = (x1', x2') = (θ,r) // note that r = x2'

x = F-1(x') ↔ x = rcos(θ) x1 = x2' cos(x1')

y = rsin(θ) x2 = x2' sin(x1')

x' = F(x) ↔ r = x2+y2 x2' = x12+x22

θ = tan-1(y/x) x1' = tan-1(x2/x1) . (1.4)

(b) The transformation is non-linear because at least one component function ( e.g., r = x2+y2 ) is not of
the form r = Ax + By. In this transformation all functions are non-linear.

(c) Here is a drawing showing the nature of this mapping:

(1.5)

The domain of x' = F(x) (in x-space on the right ) is all of R2, but the range in x'-space is the semi-infinite
vertical strip shown in gray. Imitating the language of complex variables, we can regard this gray strip as
depicting the principle branch of the multi-variable function x' = F(x). Other branches are obtained by
shifting the gray rectangle left or right by multiples of 2π. Still other branches are obtained by taking the
other branch of the real function r = x2+y2 which produces down-facing strips. The principle branch
plus all the other branches then fill up the E2 of x'-space, but we care only about the principle branch
range shown in gray (this strip could be any strip of width 2π, such as one from -π to π).

(d) This mapping illustrates a "problem point" involving θ = tan-1(y/x). This occurs when both x and y
are 0, indicated by the red dot on the right. The inverse mapping takes the entire red line segment into this
red origin point, so we have a lack of 1-to-1 going on here, meaning that formally the function F is not
invertible. This can be fixed up by eliminating the red line segment from the range of F, retaining only the
point at its left end. Another problem is that both the left and right vertical edges of the gray strip map
into the real axis in x-space, and that is fixed by removing the right edge. Thus, by doing a suitable

13
 Chapter 1: The Transformation F

trimming of the range, F can be made fudlly invertible. No one has ever had major problems using polar
coordinates due to these minor issues.

Example 2: Spherical coordinates (N=3)

(a) The transformation from Cartesian to spherical coordinates is given by,

x = (x1,x2,x3 ) = (x,y,z)
x' = (x1',x2',x3') = (r,θ,φ)

x = F-1(x') ↔ x = r sin(θ) cos(φ) x1 = x1'sin(x2')cos(x3')

y = r sin(θ) sin(φ) x2 = x1'sin(x2')sin(x3')
z = r cos(θ) x3 = x1'cos(x2')

x' = F(x) ↔ r= x2+y2+z2 x 1' = x12+x22+x32

θ = cos-1(z/ x2+y2+z2 ) x2' = cos-1(x3/ x12+x22+x32 )
φ = tan-1(y/x) x3' = tan-1(x2/x1) (1.6)

(b) The transformation is non-linear because at least one component function( e.g., r = x2+y2+z2 ) is not
of the form r = Ax + By + Cz. In this transformation, all three functions are non-linear.

(c) Here is a drawing showing the nature of this mapping

(1.7)

The domain of x' = F(x) (in x-space on the right) is all of R3, but the range in x'-space is the interior of an
infinitely tall rectangular solid on the left we shall call an "office building". We could regard this office
building as depicting the principle branch of the multi-variable function x' = F(x). Other branches are

14
 Chapter 1: The Transformation F

obtained by shifting the building left and right by multiples of 2π, or fore and aft by multiples of π, or by
flipping it vertically, taking the other branch of r = x2+y2+z2 . The principle branch plus all the other
branch offices then fill up the R3 of x-space, but we care only about the principle branch office building
whose walls are mostly shown in gray.

(d) This mapping illustrates some "problem points". One is that entire green office building main floor
(r=0) maps into the origin in x-space. This problem is fixed by trimming away the main floor keeping
only the origin point of the bottom face of the office building. Another problem is that the entire red line
segment (θ = 0) maps into the red point shown in x-space. This is fixed by throwing out the back wall of
the office building, retaining only a line going up the left edge of the back wall. A similar problem
happens on the front wall (θ = π, blue) and we fix it the same way: throw out the wall but maintain a thin
line which is the left edge of this front wall (this line is missing its bottom point). Thus, by doing a
suitable trimming of the range, F is made fully invertible.

Cartesian Space and Quasi-Cartesian Space

(a) Cartesian Space. For the purposes of this document, a Cartesian Space in N dimensions is "the usual"
Hilbert Space RN (or EN) in which the distance between two vectors is given by the formula

d(x,y) = Σi=1N (xi-yi)2 ⇒ [d(x+dx,x)]2 = Σi=1N (dxi)2

metric tensor = diag(1,1,1....1) (1.8)
as discussed in Chapter 4 below.
The θ-r space in the above Example 1 would be a Cartesian space if it were declared that the distance
between two points there were D'2 = (θ-θ')2 + (r-r')2, but that is not the usual intent in using that space.
As shown below, the metric tensor used there is g = diag (r2,1) and not diag(1,1).
One might argue that our Cartesian Space is in fact a Euclidean space (hence EN) having Cartesian
coordinates. A non-Cartesian space is sometimes referred to as a "curved space" (non-Euclidean) and the
coordinates in such a space as "curvilinear coordinates". An example is the θ-r space above.
With the Cartesian Space metric tensor as gC = 1 = diag(1,1....1), the above equations can be written

d2(x,y) = gCij(xi-yi)(xj-yj) and [d(x+dx,x)]2 = gCij dxi dxj ≡ (ds)2 (1.9)

metric tensor = gC = 1 = diag(1,1,1....1), gCij = δi,j

where repeated indices are implicitly summed, called the Einstein summation convention.

(b) Quasi-Cartesian Space. We now define a Quasi-Cartesian Space (not an official term) as one which
has a diagonal metric tensor G whose diagonal elements are independently +1 or -1 instead of all +1 as
with gC. In a Quasi-Cartesian Space the two equations above become

d2(x,y) = Gij(xi-yi)(xj-yj) and [d(x+dx,x)]2 = Gij dxi dxj ≡ (ds)2 (1.10)

and of course this allows for the possibility of a negative distance-squared (see Section 5.10).
Notice that G-1 = G for any distribution of the ±1's in G. As shown later, this means that that
covariant and contravariant versions of G are the same.

15
 Chapter 1: The Transformation F

The motivation for introducing this Quasi-Cartesian Space is to cover the case of special relativity
which involves 4 dimensional linear transformations with G = diag(1,-1,-1,-1).

Pictures A,B,C and D

We shall usually work with one of four different "pictures" involving transformations. In each picture the
spaces and transformations (and their associated objects) have certain names that prove useful in certain
situations :

(1.11)

The matrices R and S are associated with transformation F as described in Chapter 2 below, while G and
the g's are metric tensors.
Systems not marked Cartesian could of course be Cartesian, but we think of them as general "curved"
systems with strange metric tensors. And in general, all the full transformations might be non-linear.
The polar coordinates example above was presented in the context of Picture B. Picture B is the right
picture for studying curvilinear coordinates where for example x-space = Cartesian coordinates and x'-
space = toroidal coordinates. Picture C is useful for making statements applying to objects in curved x-
space where we don't want lots of primes floating around. Pictures A and D are appropriate for
consideration of general transformations, as well as linear ones like rotations and Lorentz transformations.
In Chapters 9-14 Picture M&S (Moon & Spencer) is introduced for the special purpose of displaying the
differential operator expressions. This is Picture B with x'→ u and g'→g on the left side.
The entire rest of this Section uses the Picture B context.

Coordinate Lines

Suppose in x'-space one varies a single coordinate, say x'i, keeping all the other coordinates fixed. In x'-
space the locus of points thus created is just a straight line parallel to the x'i axis, or for a principle branch
situation like that of the above examples, a straight line segment. When such a straight line or segment is

16
 Chapter 1: The Transformation F

mapped into x-space, the result is a curve known as a coordinate line. A coordinate line is associated
with a specific x'-space coordinate x'i, so one might refer to the " x'i -coordinate line", x'i being a label.

In N dimensions, a point x in x-space lies on a unique set of N coordinate lines with respect to a
transformation F. Remember that each such line is associated with one of the x'i coordinates. In x'-space,
a point x' lies on a unique intersection of straight lines or segments, and then this all gets mapped into x-
space where point x = F-1(x') then lies on a unique intersection of coordinate lines.
For example, in spherical coordinates we start with some (x,y,z) in x-space and compute the x'i =
(r,θ,φ) in x'-space. Our point x in x-space then lies on the r-coordinate line whose label is r, it lies on the
θ-coordinate line whose label is θ, and it lies on the φ-coordinate line whose label is φ (see below).

In general a coordinate "line" is some non-planar curve in N-dimensional x-space, meaning that a
coordinate line might not lie on an N-1 dimensional plane. In the 2D polar coordinates example below,
the red coordinate line does not lie on a 1-dimensional plane (line). In the next example of 3D spherical
coordinates, it happens that every coordinate line does lie on a 2-dimensional plane. But in ellipsoidal
coordinates, another 3D orthogonal system, every coordinate line does not lie on a 2-dimensional plane.

Some authors refer to coordinate lines as level curves, especially in two dimensions mapping the real and
imaginary part of analytic functions w = f(z) (Ahlfors p 89).

Example 1: Polar coordinates, coordinate lines

Here are some coordinate lines for our prototype N=2 non-linear transformation, Cartesian to polar
coordinates:

(1.12)

The red circle is a θ-coordinate line, and the blue ray is an r-coordinate line

17
 Chapter 1: The Transformation F

Example 2: Spherical coordinates, coordinate lines

These coordinate lines are generated exactly as described above. In x'-space one holds two coordinates
fixed while allowing one to vary. The locus in x'-space is a line segment or a half line (in the case of
varying r). In x-space, the corresponding coordinate lines are as shown.

(1.13)

The green coordinate line is a θ-coordinate line, since only θ is varying.

The red coordinate line is an r-coordinate line, since only r is varying.
The blue coordinate line is a φ-coordinate line, since only φ is varying.

The point x indicated by a black dot in x-space lies on the unique set of coordinates lines shown.

Appendix C gives an example of coordinate lines for a non-orthogonal 2D coordinate system.

Level Surfaces

(a) Suppose in x'-space one fixes one coordinate, say x'i, and varies all the other coordinates. In x'-space
the locus of points thus created is just an (N-1 dimensional) plane perpendicular to the xi axis, or for a
principle branch situation like that above, a rectangle or half strip in the case of r. Mapping this planar
surface in x'-space into x-space produces a surface in x-space (of dimension N-1) called a level surface.
The equations of the N different xi level surface types are

a'i(n) = Fi(x1, x2.....xN) i = 1,2...N (1.14)

18
 Chapter 1: The Transformation F

where a'i(n) is some constant value selected for fixed coordinate x'i. By taking some set of closely
spaced values for this constant, { a'i(1), a'i(2).....}, one obtains a family of level surfaces all of the same
general shape which are closely spaced. For some different value of i, the shapes of such a family of level
surfaces will in general be different. In general if f(x1, x2.....xN) = k, the set of points x which make this
equation true for some fixed k is called a level set, so a level set is a surface of dimension N-1. Thus, all
our level curves are also level sets.

In the polar coordinates example, since there are only 2 coordinates, there is no distinction between a
level surface and a coordinate line.
In the spherical coordinates example, there is a distinction.
If one fixes r and varies θ and φ over their horizontal rectangle inside the office building, the level
surface in x-space is a sphere.
If one fixes θ and varies r and φ over a left-right vertical strip inside the office building, the level
surface in x-space is a sphere is a polar cone
If one fixes φ and varies r and θ over a fore-aft vertical strip inside the office building, the level
surface in x-space is a half plane at azimuth φ.

(b) In N dimensions there will be N level surfaces in x-space, each formed by holding some x'i fixed. The
intersection of N-1 level surfaces (omitting say the x3' level surface) will have all of the x'i fixed except
x'3. But this describes the x'3 coordinate line. Thus, each coordinate line can be considered as the
intersection of the N-1 level surfaces associated with the other coordinates. One can see this happening on
the spherical coordinates example:

The green coordinate line is the intersection of two level surfaces: half-plane and sphere.
The red coordinate line is the intersection of two level surfaces: half-plane and cone.
The blue coordinate line is the intersection of two level surfaces: sphere and cone.

19
 Chapter 2: Linear Transformations R and S

2. Linear Local Transformations associated with F : scalars and two kinds of vectors

2.1 Linear Local Transformations

We now shift to the Picture A context, where x-space is not necessarily Cartesian,

(2.1.1)

Consider again the possibly non-linear transformation x' = F(x) mapping F: RN→ RN. Imagine a very
small neighborhood around the point x in x-space, a "ball" around x. Where the mapping is continuous in
both directions, one expects a tiny x-space ball around x to map into a tiny x'-space ball around x' and
vice versa. Here is a picture of this situation,

(2.1.2)

where everything in one picture is the mapping of the corresponding thing in the other picture.
In particular, we show a small vector in x-space called dx which maps into a small vector in x'-space
called dx'. Since F was assumed invertible, it must be invertible locally in these two balls. That is, given a
dx above, one can determine dx', and vice versa. (Anticipating a few lines below, this means that the
matrices S and R will be invertible so neither can have zero determinant.)
How are these two differential vectors related? For a linear approximation,

x'i + dx'i = Fi(x + dx ) ≈ Fi(x) + Σk( ∂Fi(x)/∂xk) dxk

⇒ dx'i = Σk( ∂Fi(x)/∂xk) dxk . (2.1.3)

The last line shows an equals sign in the limit that dxk is a vanishing differential. Since Fi(x) = x'i ,

dx'i = Σk(∂x'i/∂xk) dxk = Σk Rik dxk where Rik ≡ (∂x'i/∂xk) . (2.1.4)

Doing the same operation in the other direction gives

dxi = Σk( ∂xi/∂x'k) dx'k = Σk Sik dx'k where Sik ≡ (∂xi/∂x'k) . (2.1.5)

20
 Chapter 2: Linear Transformations R and S

One can regard Rik and Sik as elements of NxN matrices R and S. In vector notation then,

dx' = R(x) dx Rik(x) ≡ (∂x'i/∂xk) R = S-1 // dx'i = Rij dxj

dx = S(x') dx' Sik(x') ≡ (∂xi/∂x'k) S = R-1 // dxi = Sij dx'j . (2.1.6)

It is obvious that matrices R and S are inverses of each other, just staring at the above two vector
equations. One can verify this fact from the definitions of R and S using the chain rule

∂x'i ∂xk ∂x'i

(RS)ij = Σk RikSkj = Σk (∂x'i/∂xk) (∂xk/∂x'j) = Σk ∂x ∂x' = ∂x' = δi,j . (2.1.7)
k j j

We could get rid of one of these matrices right now, perhaps keeping R and replacing S = R-1, but
keeping both simplifies expressions encountered later, so for now both are kept.
The letter R does not imply that matrix R is a rotation matrix, although it could be. According to the
polar decomposition theorem (Lai p 110), any matrix R (detR ≠ 0) can be uniquely written in the form
R = RU = VR where R is a rotation matrix (the same one in RU and VR) and U and V are symmetric
positive definite matrices (called right and left stretch tensors) related by U = RTVR. Matrix S could of
course be written in a similar manner.
Matrices R(x) and S(x') are in general functions of a point in space x' = F(x). As one moves around in
space, all the elements of matrices R and S are likely to change. So R and S represent point-dependent
linear transformations which are valid for the differentials shown.
One might wonder at this point how the vector dx is related to its components dxi and the same
question for dx'i and dx'i. As will be shown later in (6.6.9) and (6.6.15),

dx = Σndxn un where the un are x-space axis-aligned basis vectors of the form u1 = (1,0,0,..0)
dx' = Σndx'n e'n where the e'n are x'-space axis-aligned basis vectors of the form e'n = (1,0,0,..0)
(2.1.8)

If x-space and x'-space were Cartesian, one could write un = ^ n and e'n = ^ n', but in general the un and e'n
vectors do not have (covariant) unit length, as shown later in (6.5.3) and (6.4.1).
The reader familiar with covariant "up and down" indices will notice that all indices are peacefully
sitting "down" in the presentation so far (subscripts, no superscripts). As we carry out our various
developmental tasks, that is where all indices shall remain until Chapter 7, whereupon they will start
frantically bobbing up and down, seemingly at will. [ Since rules are made to be violated, we have
violated this one in some examples below where non-standard notation would be hard to swallow. ]
Are there any "useful objects" that can be constructed from differentials dx and which might then
transform according by R or S? The answer is yes, but first we discuss scalars.

2.2 Scalars

A quantity is a scalar with respect to transformation F if it is the same in both spaces. Thus, any constant
like π would be a scalar under any transformation. The mass m of a potato would be a constant under
transformations that are rotations or translations. A function of space φ(x) is a "field" and it would be a
"scalar field" if φ'(x') = φ(x). For example, temperature would be a scalar field under rotations. Notice

21
 Chapter 2: Linear Transformations R and S

that φ is evaluated at x, while φ' is evaluated at x' = F(x). As noted in Section 2.12, one could be more
precise by referring to the objects described here as a "tensorial scalar" and a "tensorial scalar field".

2.3 Contravariant vectors

If x and x' are spatial coordinates (time t is not one of the xi), then consider the familiar velocity vector,

vi = dxi/dt ⇒ v = dx/dt . (2.3.1)

Since dt transforms as a constant (scalar) under our selected transformation type, it seems pretty clear that
velocity in x'-space can be related to velocity in x-space using the dx' = R(x) dx rule above, so

v' = R(x) v . (2.3.2)

Even though the matrix R(x) changes as we move around, this linear transformation R is valid at any
point x when applied to velocity. Momentum p = mv would work the same way, since mass m is a scalar
(Newtonian mechanics).
In contrast, unless R(x) is a constant in space (meaning from (2.8.7) that F(x) is linear), x' ≠ R(x) x,
so in general x itself is not a contravariant vector although dx is.
Any vector that transforms according to V' = R(x)V with respect to a transformation F (such as
Newtonian velocity and momentum with respect to rotations) is called a contravariant vector.

2.4 Covariant vectors

Much of physics is described by differential equations involving the gradient operator ( the reason for the
overbar is explained in the next Section)

¯i = ¯
∇ ∂ i = ∂/∂xi (2.4.1)

which involves an "upside down" differential. Here is how this operator transforms going from x-space to
x'-space, again according to the chain rule (implied sum on k, T means transpose)

∂ ∂xk ∂
¯ 'i = ¯
∇ ∂ 'i = ∂x' = ∂x' ∂x ∂ k = STik ¯
= Ski¯ ∂ k = STik ∇
¯ k = (ST ∇
¯ )i
i i k

¯ ' = ST ∇
⇒ ∇ ¯. (2.4.2)

¯ as acting on a scalar field φ(x) = φ'(x'), and then the above becomes
One can think of ∇

∂x
¯ 'i φ'(x') = ∂ φ'(x') = k ∂ φ(x) = STik ∇
∇ ¯ k φ(x)
∂x'i ∂x'i ∂xk

¯ 'φ'(x') = ST ∇
⇒ ∇ ¯ φ(x) . (2.4.3)

22
 Chapter 2: Linear Transformations R and S

Since the differential is "upside down", one might expect ∇ ¯ to transform according to S = R-1 instead of
R, but it is really ST that does the job. One could write ∇
¯' = ∇
¯ S in terms of row vectors.
T ¯ are called covariant
Vectors that transform according to V' = S (x) V such as the gradient operator ∇
vectors with respect to transformation F.
An example of a covariant vector is the electrostatic electric field obtained from the potential Φ

¯Φ
Ē = - ∇ Ēi = - ¯
∂ iΦ = - ∂Φ/∂xi (2.4.4)

2.5 Bar notation

In order to distinguish a contravariant from a covariant vector, we shall (for a while) adopt this bar
convention: contravariant vectors shall be written V with components Vi and covariant vectors shall be
written V̄ with components V̄i. This is why overbars were placed on ∇ ¯ and ¯ ∂ i and Ē in the previous
Section. We call this our "developmental notation", as distinct from the Standard Notation introduced in
Chapter 7.
The transformation rules for the two vector types can now be written this way:

V' = R V contravariant Rik(x) ≡ (∂x'i/∂xk) R = S-1

V̄' = ST V̄ covariant Sik(x') ≡ (∂xi/∂x'k) = STki(x') (2.5.1)

One could imagine replacing S with some QT to make the second equation more like the first, but of
course then RQT = 1 instead of RS = 1. In the Standard Notation, where there are four versions of the
matrix R, we shall see in Section 7.5 that R → Rij and S → Sij = Rji and S can be removed from the
picture altogether.
∂x'i
It is possible to do without either R or S and just write out all partial derivatives like ∂xk in place of
Rik. Many authors do this, including Weinberg. We feel that the forms Rik and later Rik are much more
compact (3 symbols in place of 8) and reveal more clearly the matrix (but not tensor) nature of R and the
fact that it is a linear transformation that locally approximates the full x' = F(x). Later we shall have long
strings of R objects which become extremely cluttered when derivatives are fully written out, see for
example Section F.2 and following. Finally, Rik is a easy to type!

2.6 Origin of the names contravariant and covariant

A justification of the terms covariant and contravariant is presented in Section 7.14, since the idea is more
easily presented there than here.
It seems that these terms were first used in 1851 (a half century before special relativity) in a paper
(see Refs.) by J.J. Sylvester of Sylvester's Law of Inertia fame. Sylvester uses the words covariant and
contravariant to describe the relations between a pair of "transformations". In much simpler notation than
he uses, if those "transformations" (functions) are F(x) and G(x) and if A is an 3x3 matrix, then

23
 Chapter 2: Linear Transformations R and S

the pair F(Ax) and G(Ax) are said to be covariant (or concurrent)
the pair F(Ax) and G(A-1x) are said to be contravariant (or reciprocal)

The idea is that in comparing the way two things transform, if they both move the same way, then it is
covariant, and if they move in opposite directions it is contravariant. In Section 7.14 this idea is applied to
the transformation of two objects, where one object is the component of a vector like Vn and the other
object is a basis vector onto which a vector is expanded. The connection is a bit distant, but the
underlying concept carries through.
Notations like y = F(Ax) would have mystified Sylvester in 1851, although in this same paper he
introduced two-dimensional arrays of letters and referred to them as "matrices". According to a web piece
by John Aldrich of the University of Southampton, J.W. Gibbs in 1881 was the first person to use a single
letter to represent a vector (he used Greek letters). It was not until 1901 when his student E.B. Wilson
published Gibb's lectures in a Vector Analysis book that the idea was propagated to a wider circle. Wilson
converted those Greek letters to bolded ones,

(2.6.1)

The Wilson/Gibbs book was reprinted seven times, the last being 1943. In 1960 it continued as a Dover
book and is now available online as a public domain document (above from page 4).

2.7 Other vector types?

Are there any other kinds of vectors with respect to a transformation F? There might be, but only the two
types mentioned above are of interest to us in this document. They are both called rank-1 tensors, and
there are no other rank-1 tensor types in "tensor analysis" (for rank-n tensors, see Section 7.10). Some
authors refer to the rank of a tensor as the order of a tensor, and we shall sometimes use that term.
In the Standard Notation introduced later, where contravariant vector components are written with
indices up and covariant vectors with indices down, and where the notation is so slick and smooth and
automatic, one sometimes imagines there are two kinds of vectors because there are two places to put
indices, up and down. It is of course the other way around: the up/down notation was adopted because
there are two rank-1 tensor types.
Two particular (linear) transformation types of interest are rotations and Lorentz transformations,
each of which has a certain number of continuous parameters (3 and 6). As the parameters are allowed to
vary over their ranges, the set of transformations can be viewed as elements of a continuous group
( SO(3) and SO(3,1) ). Each of these groups has exactly one "vector representation" ( "1" and
"(1/2)⊕(1/2)" ). One should not imagine that somehow the "two-ness" of vector types under general
transformations F is connected to there being two vector representations of some particular group. It

24
 Chapter 2: Linear Transformations R and S

happens that the Lorentz group does have two "spinor representations" (1/2)⊕0 and 0⊕(1/2), but this has
nothing at all to do with our general notion of two kinds of vectors. This subject is discussed in more
detail in Section 5.14.

2.8 Linear transformations

First, consider a scalar function f : Rn → R. The function f(x) is linear iff

f(x+y) = f(x) + f(y) for all x and y in Rn

f(ax) = af(x) for all x in Rn and for all a in R . (2.8.1)

These conditions imply that the function f(x) must have the form,

f(x) = Σi ai xi . where ai are constants independent of x . (2.8.2)

One way to reach this conclusion is to assume a Taylor expansion for f(x) about x = 0,

f(x) = f(0) + Σi[∂if]x=0xi + Σij[∂ijf]x=0xixj + higher terms . (2.8.3)

The (2.8.1) requirement that f(2x) = 2f(x) eliminates the quadratic and higher terms and for x = 0
eliminates the first term f(0) since f(0) = 2f(0), resulting in an expression of the form (2.8.2).

When this same discussion is applied to the components Fi(x) of the function F : Rn→Rn one concludes
that the form of x' = F(x) for linear F is given by

x'i = Fi(x) = ΣjFijxj where Fij are constants independent of x . (2.8.4)

Applied to the differential vector dx this says

dx'i = Σj Fij dxj . (2.8.5)

Comparison with (2.1.4) shows that Fij = Rik so

F linear ⇒ R = F and S = F-1 . (2.8.6)

We then arrive at this miniature theorem and its converse:

Theorem:
(a) If R(x) = R does not vary with x, then dx' = Rdx ⇒ x' = Rx so x' = F(x) is a linear
transformation, namely F(x) = Rx .
(b) If x' = F(x) is a linear transformation, then F(x) = Rx where R is a matrix of constants which do
not vary with x. (2.8.7)

This is the situation for global rotations and Lorentz transformations.

25
 Chapter 2: Linear Transformations R and S

2.9 Vectors that are contravariant by definition

A contravariant vector has been defined above as any N-tuple which transforms the same way that dx
transforms with respect to F, namely, dx' = R(x) dx. One might state this as

{ dx', dx } dx' = R(x) dx contravariant vector . (2.9.1)

Suppose we start with an arbitrary N-tuple V and simply define V' ≡ RV. One would have to conclude
that the pair { V', V } transforms as a contravariant vector.

{ V', V } V' ≡ R(x)V contravariant vector . (2.9.2)

Conversely, one could start with some given V' and define V ≡ S(x)V' (recall S = R-1), and again one
would conclude that { V', V } represents a vector that transforms as a contravariant vector.
We refer to either process as producing a vector which is "contravariant by definition". Creating a
contravariant vector in this fashion is a fine thing to do, as long as the defined vector does not conflict
with something that already exists.

Example 1: We know that if F is non-linear, the vector x does not transform as a contravariant vector,
because x' = R(x)x is not true, where x' = F(x). If we start with x and try to force {x', x} to be
"contravariant by definition" by defining x' ≡ R(x) x, this x' conflicts with the existing x' = F(x), so the
method of contravariant by definition is unacceptable. (2.9.3)

Example 2: As another example, consider an N-tuple in x'-space of three masses V' = (m1,m2,m3). The
transformation is taken in this example to be regular rotations. Since masses are rotational scalars with
respect to such rotations, we know that in an x-space rotated frame of reference we would find V =
(m1,m2,m3). We could attempt to set up { V', V } as a vector that is "contravariant by definition" by
defining V ≡ SV', but this conflicts with the existing fact that V = (m1,m2,m3), so the method of
contravariant by definition is again unacceptable. (2.9.4)

Example 3: This time F is a general transformation and we start with V' = e'n which are a set of axis-
aligned basis vectors in x'-space. We define vectors V = en according to en ≡ Se'n. Then { e'n, en } form a
vector which is "contravariant by definition" and e'n = R en (R = S-1). Since the newly defined vector en
does not conflict with some already-existing vector in x-space, the method of contravariant by definition
in this example is acceptable. This is exactly what is done in the next Section with the tangent base
vectors en. (2.9.5)

2.10 Vector Fields

We considered above vectors like position x (and dx) and velocity v and the vector operator ∇ ¯ , and we
referred to a generic vector as V. Many vectors of interest (in fact, most) are functions of x, which is to
say, they are vector fields. Examples are the electric and magnetic fields E(x) and B(x), or the average
velocity of a small region of fluid V(x) or a current density J(x). Another example is the transformation
F(x).

26
 Chapter 2: Linear Transformations R and S

We already mentioned scalar fields, such as temperature T(x) or electrostatic potential Φ(x). The way
a scalar temperature field transforms going from x-space to x'-space is this

T '(x') = T(x) where x' = F(x) . (2.10.1)

If the transformation is a 3D rotation from frame S to frame S', then T ' is the temperature measured in
frame S' at point x' and T is the temperature measured at the corresponding point x in frame S and of
course there is only one temperature at that point so the numbers are equal. In x'-space one needs the
prime on T ' because the functional form (how T ' depends on the x'i) is not the same as that of T (how T
depends on the xi). For example, if transformation F is from 2D Cartesian to polar coordinates, then

T '(r,θ) = T(x,y) = T(rcosθ,rsinθ) ≠ T(r,θ) . (2.10.2)

Contravariant and covariant vector fields transform as described above, but now one must show the
argument for each field in its own space, and again x' = F(x) :

V'(x') = R V(x) contravariant Rik(x) ≡ (∂x'i/∂xk) R = S-1

V̄'(x') = ST V̄(x) covariant Sik(x') ≡ (∂xi/∂x'k) = STki(x') . (2.10.3)

Similar transformation rules apply to tensors of any rank. For example (as we shall see in Chapter 5) the
metric tensor gab (developmental notation) is a rank-2 contravariant tensor field and the transformation
rule is this (implied summation on repeated indices),

g'ab(x') = Raa'Rbb'ga'b'(x) or g'ab = Raa'Rbb'ga'b' (2.10.4)

Often the coordinate dependence of g is suppressed, just as it is for R and S, as shown on the right above.

Jumping momentarily into Standard Notation, in special relativity one has x'μ = Λμνxν where F = R = Λ is
a linear transformation, and one would then specify the transformation of a contravariant vector field as

V'μ(x'α) = Λμν Vν(xα) x'μ = Λμνxν (2.10.5)

Comment. The word field in "vector field V(x)" is unrelated to the word field being an algebraic object
with • and + properties, such as the field of the real numbers or the finite field {0,1}. However, the
components of V and x in V(x) are normally elements of the real number field.

2.11 Names and symbols

The matrix Rik(x) = (∂x'i/∂xk) is called the Jacobian matrix for the transformation x' = F(x) , while the
matrix Sik(x') = (∂xi/∂x'k) is then the Jacobian matrix of the inverse transformation x = F-1(x'). The
determinant of the Jacobian matrix S will be shown in Section 8.6 to have a certain significance, and that
determinant is called "the Jacobian" = det(S(x')) ≡ J(x').
Matrix R is sometimes called "the differential dF" of transformation x' = F(x) and is. Matrix S is then
the differential of the inverse transformation x = F-1(x').

27
 Chapter 2: Linear Transformations R and S

The author has anguished over what names to give the matrices R and S = R-1. One option was to use
R = L, where L stands for the fact that this matrix is describing a Local coordinate system at point x, or a
Linearized transformation. But L is commonly used for differential operators and angular momentum, so
that got rejected. R is often called Λ in special relativity, but why go Greek so early? Another option is to
use R = J for Jacobian matrix, but J looks too much like "an integer" or angular momentum or "the
Jacobian". T for Transformation might have been confused with the full transformation F, or Cauchy
stress T. Our chosen notation R makes one think perhaps R is a Rotation, but that won't in general be the
case. For the moment we will continue to use R and S, where recall RS = 1. We shall refer to R simply as
"the R matrix for transformation F".
The fact that vectors are processed by NxN matrices R and S puts that part of the subject into the field
of linear algebra, and that may be the origin of the name tensor algebra as a generalization of this idea
(tensors as objects of direct product algebras). Of course the differential calculus aspect of the subject is
already highly visible, there are ∂ symbols everywhere (hence the name tensor calculus).

2.12 Definition of the words "scalar", "vector" and "tensor"

In Section 2.2 a "scalar" was defined as something that is invariant under some transformation F, and this
was identified with a "rank-0 tensor". Similarly, a "vector" is either a contravariant vector or a covariant
vector and both of these are "rank-1 tensors". In Section 5.6 certain "rank-2" tensors will appear -- they
are matrices that transform in a certain way under a transformation F. In Section 7.10 tensors of rank-n
will appear, and these are objects with n indices which transform in a certain manner under F.
To be more precise and to provide protection against the vagaries of "the literature", these objects
probably should have been defined with the word "tensorial" in front of them.

"tensorial scalar" ≡ rank-0 tensor with respect to some transformation F

"tensorial vector" ≡ rank-1 tensor with respect to some transformation F
"tensorial tensor" ≡ rank-n tensor with respect to some transformation F (2.12.1)

As has been emphasized several times, a "tensorial tensor" is linked to a particular underlying
transformation F, and one should really use the more precise term "tensorial tensor under F".
In this paper, we generally omit the word "tensorial" when discussing the above objects. This brings
us into conflict with the following definitions which are often used: (Here, we use the term "expression"
to indicate a number, a variable, or some combination of same.)

• A "scalar" is a single expression, a 1-tuple. No invariance under any transformation is implied.

• A "vector" is an N-tuple of expressions. No transformation rule is implied.

• A "second order tensor" is a matrix of expressions. No transformation rule is implied. (2.12.2)

• A "tensor" is an object with n indices, n = 2,3,4... which includes the previous item. A tensor is
therefore a collection of expressions which are labeled by n indices each of which goes 1 to N. No
transformation rule is implied.

To these definitions we can add another list:

28
 Chapter 2: Linear Transformations R and S

• A "scalar field" is a single function of x ( the x-space coordinates). No implication of invariance.

• A "vector field" is an N-tuple of functions of x -- an N-tuple of scalar fields. No transform implied.

• A "tensor field" of order n is a set of Nn scalar functions, for example, Tabc...(x). Same. (2.12.3)

In any discussion which includes relativity (special or general), the words scalar, vector and tensor would
always imply the tensorial definitions of these words. Continuum mechanics, however, seems to use the
above alternate list of definitions, so that any matrix is called a tensor. Usually such matrices are
functions of space and should be called tensor fields, but everybody knows what is meant.
In Section 7.15 we shall discuss the notion of an equation being covariant, which means it has the
exact same form in different frames of reference which are related by a transformation. For example, one
might have F = ma in frame S, and F' = m'a' in frame S', where these frames are related by a static
rotation. F and a are tensorial vectors with respect to this rotation, and m and m' are tensorial scalars, and
m = m' for that reason. Both sides of F = ma transform as tensorial vectors. Since rotations are an
invariance of Newtonian physics, any valid equation of motion must be "covariant", and this applies of
course to particle, rigid body and continuum mechanics.
In the latter field, continuum mechanics, one naturally seeks out model equations which are covariant.
In order to do this properly, one must know which tensors are tensorial tensors, and which tensors are just
tensors with either no transformation rule, or some transformation rule that does not match the tensor.
Continuum mechanics has evolved special words to handle this situation. If a tensor is a tensorial tensor,
it is said to be objective, or indifferent. In continuum mechanics an equation which is covariant is said to
be frame-indifferent.

In this document we shall follow the time-honored tradition of being inconsistent in our use of the words
scalar, vector and tensor, but the reader is now at least warned. To further promote that tradition, we shall
sometimes refer to tensorial tensors as "true" tensors.

The notion of tensor densities described in Appendix D further complicates the nomenclature. One can
have scalar densities and vector densities of various weights, for example.

Appendix K explores a few commonly used tensors in continuum mechanics and determines which of
these tensors actually transform as tensors (are objective), and which tensors do not transform as tensors
(are non-objective).

Comment on "rank". As noted in Section 2.7, some authors refer to a rank-n tensor as an order-n tensor.
The word rank has another common use which is unrelated to tensor rank. The rank of a square matrix is
the number of linearly independent rows or columns. If an NxN matrix M has rank less than N, then
det(M) = 0. Since our R and S matrices do not have zero determinant, they are both of full rank N. The
same is true of the metric tensors of Chapter 5 below. Since g ≡ det(ḡ ij) ≠ 0, g has full rank N.

29
 Chapter 3: Tangent Base Vectors

3. Tangent Base Vectors en and Inverse Tangent Base Vectors u'n

3.1 Differential Displacements

This entire Section is in the context of Picture A,

(3.1.1)

In Fig (2.1.2) above showing dx and dx', one has much freedom to "try out" different differential vectors.
For any dx one picks at point x, one gets some dx' according to dx' = R(x) dx. Consider this slightly
enhanced version of that figure (red curves added)

(3.1.2)

The point x in x-space (right side) can be regarded as lying on some arbitrary 1-dimensional curve in RN
shown on the right in red. Select dx to be the tangent to this curve at point x. That curve will then map
into some (probably very different) curve in x'-space which passes through the point x'. The tangent to
this curve at the point x' must be dx' = R(x) dx. A similar statement can be made starting instead with an
arbitrary curve in x'-space. The tangent dx' there then maps into dx = S(x') dx' in x-space.
The curves are in N-dimensional space and are in general non-planar and the tangents are of course N
dimensional tangents, so this 2D picture is mildly misleading.

We now specialize such that the red curve on the left is a straight line parallel to an x'-space axis, which
means the curve on the right is a coordinate line,

30
 Chapter 3: Tangent Base Vectors

(3.1.3)

Admittedly the drawing does not strongly suggest that the red line segment on the left is parallel to an
axis in x'-space, but since those axes are not drawn, one cannot complain too strenuously.

3.2 Definition of the en ; the en are the columns of S

First, define a set of N basis vectors in x'-space which point along the positive axes of x'-space,

e'n , n = 1,2...N (e'n)i = δn,i e'1 = (1,0,0...) etc . (3.2.1)

Assume that the dx' arrow above points in this e'n direction so that

dx' = e'n dx'n // no implied sum on n (3.2.2)

where dx'n is a positive differential variation of coordinate x'n along the e'n axis in x'-space. The
corresponding dx in x-space will be,

dx = S dx' = S [e'n dx'n] = [ Se'n] dx'n ≡ en dx'n (3.2.3)

where this last equality serves as the definition of en ,

en ≡ Se'n . (3.2.4)

Vector en = en(x) points along dx in x-space and is tangent to the x'n- coordinate line there at point x.
This vector en is generally not a unit vector, hence no hat ^ . Writing out the ith component of (3.2.4),

(en)i = Σj Sij (e'n)j = Σj Sij δn,j = Sin (3.2.5)

so that, with (2.1.5),

(en)i = Sin = ∂xi/∂x'n or en = ∂x/∂x'n = ∂'nx . (3.2.6)

This fact that (en)i = Sin says that the vectors en are the columns of the matrix S:

S = [e1, e2, e3 .... eN ] matrix = N columns (3.2.7)

31
 Chapter 3: Tangent Base Vectors

We shall call these en vectors the tangent base vectors. The vectors exist in x-space and point along the
various coordinate lines that pass through a point x.
If the points on the x'n-coordinate line were labeled with the values of x'n from which they came, one
would find that en points in the direction in which those labels increase.
As one moves from x to some nearby point, the tangent base vectors all change slightly because in
general S = S(x'(x)) and the en = en(x) are the columns of S. Any set of basis vectors which depends on x
in this way is called a local basis. In contrast, the corresponding x'-basis e'n shown above with (e'n)i =
δn,i is a global basis in x'-space since it is the same at any point x' in x'-space.
Since det(S) ≠ 0 due to our assumption that F is invertible, the tangent base vectors are linearly
independent and provide a basis for EN.
One can of course normalize each of the en to be a unit vector ^ e n according to ^
e n = en/ |en|.
Here is a traditional N=3 picture showing the tangent base vectors pointing along three generic
coordinate lines in x-space all of which pass through the point x:

(3.2.8)

Comment on notation. Some authors refer to our en as gn or Rn or other. Later it will be shown that
en•em = ḡ'nm where ḡ'nm is the covariant metric tensor for x'-space, so admittedly this provides a
reasonable argument for using gn so that gn•gm = ḡ'nm. But then the primes don't match which is
confusing: the gn are vectors in x-space, while ḡ' is a metric tensor in x'-space. We shall be using yet
another g in the form g = det(ḡ nm) and a corresponding g'. Due to this proliferation of g objects, we stick
with en, the notation used by Margenau and Murphy (p 193). A g-oriented reader can replace e → g as
needed anywhere in this document. As for unit vector versions of the en, we use the notation ^ e n ≡ en/|en|.
Morse and Feshbach use an for this purpose (Vol I p 22). A g-person might use ^ gn .
A related issue is what symbols to use for the "usual" basis vectors in Cartesian x-space. As noted
above in (2.1.8), we are using un with (un)i = δn,i as "axis-aligned basis vectors" in x-space. If ḡ = 1 for
x-space, then these are the usual Cartesian unit vectors (see (6.5.3) below that un • um = ḡ nm). Many
authors use the notation en for these vectors which then conflicts with our use of en as the tangent base
vectors. Morse and Feshbach use the symbols i, j, k for our Cartesian u , u , u . Other authors use ^
1 2 3 1, ^
2, ^
3
so then un = ^
n.

32
 Chapter 3: Tangent Base Vectors

Often the notation en is used by authors to represent some generic arbitrary set of basis vectors. For
this purpose, we shall use the notation bn.

3.3 en as a contravariant vector

The situation described above is this,

dx' = e'n dx'n x'-space (3.2.2) // no implied sum on n

dx = en dx'n x-space (3.2.3) // no implied sum on n (3.3.1)

and the full transformation F maps dx into dx'. Since dx is a contravariant vector, the linear
transformation R also maps dx into dx'. Thus

dx' = R(x) dx (2.1.6)

so
[e'n dx'n] = R(x) [en dx'n] (3.3.1)
so
e'n = R(x) en . (3.3.2)

We can regard the last line as a statement that the vector en transforms as a contravariant vector under F.
Written out in components one gets

(e'n)i = ΣjRij (en)j ⇒ δn,i= ΣjRijSjn (3.3.3)

recovering the fact that RS = 1. This is an example of a vector being "contravariant by definition", as
discussed in (2.9.5).

The two expansions (3.3.2) and (3.2.4) are easy to verify by showing that the components of both sides
are the same:

e'n ≡ Ren = Σi Rin ei since (e'n)j = Σi Rin (ei)j = Σi Rin Sji = (SR)jn = δj,n = (e'n)j

en ≡ Se'n = Σi Sin e'i since (en)j = Σi Sin (e'i)j = Σi Sin δi,j = Sjn = (en)j (3.3.4)

3.4 A semantic question: unit vectors

Above it was noted that e'1 = (1,0,0....). Should this be called "a unit vector" ? It will be seen in (6.2.7)
that in fact |e'1| = ḡ'11 ≠ 1 where ḡ' is the covariant metric tensor in x'-space, and |e'1| is the covariant
length of e'1. So e'n is a unit vector in the sense that it has a single 1 in its column vector definition, but it
is not a unit vector in the sense that it does not (in general) have unit magnitude (it would if x'-space were
Cartesian with g'=1).We take the magnitude = 1 requirement as the proper definition of a unit vector. For
this reason, we refer to the e'n in x'-space as just "axis-aligned basis vectors" and they have no "hats".
One wonders how such a vector should be depicted in a drawing, see Example 1 (b) below and also
Section C.5

33
 Chapter 3: Tangent Base Vectors

Example 1: Polar coordinates, tangent base vectors

(a) The first step is to compute the matrix Sik(x') ≡ (∂xi/∂x'k) from the inverse equations:

x = (x1, x2 ) = (x,y)
x' = (x1', x2') = (θ,r) // note that r chosen as the second variable x2'

x = F-1(x') ↔ x = rcos(θ) x1 = x2' cos(x1')

y = rsin(θ) x2 = x2' sin(x1') (1.4)
So
S11 = (∂x/∂θ) = -rsinθ
S12 = (∂x/∂r) = cosθ Sik ≡ ( ∂xi/∂x'k)
S21 = (∂y/∂θ) = rcosθ
S22 = (∂y/∂r) = sinθ
(3.4.1)
-rsinθ cosθ -sinθ/r cosθ/r
S = ⎛⎝ rcosθ sinθ ⎞⎠ // det(S) = -r R = S-1 = ⎛⎝ cos(θ) sinθ ⎞⎠ .

[ Note: The above S and R are stated for the ordering 1,2 = θ,r . For the more usual ordering 1,2 = r,θ the
colums of S should be swapped, and the rows of R should be swapped. In the usual ordering, det(S) = +r.]

The tangent base vectors en can be read off as the columns of S according to (3.2.7),

e1 = r(-sinθ,cosθ) = eθ = r ^
eθ // = r ^
θ
e2 = (cosθ,sinθ) = er = ^
er . // = ^ r (3.4.2)

Notice that eθ in this case is not a unit vector. Below is a properly scaled drawing showing the location of
the two x'-space basis vectors on the left, and the two tangent base vectors on the right. As just shown, the
length of er is 1, while the length of eθ is 2.

(3.4.3)
r and eθ = r ^
The tangent base vectors are fairly familiar animals, since er = ^ θ in usual parlance. If one
moves radially outward from point x, the er base vector stays the same, but eθ grows longer. If one moves

34
 Chapter 3: Tangent Base Vectors

azimuthally from x to some larger angle θ+Δθ, both vectors stay the same length but they rotate together
staying perpendicular.

(b) This is a good place to point out that vectors drawn in a non-Cartesian space can have magnitudes
which do not equal the length of the drawn arrows. The "graphical arrow length" of a vector v is (vx2 +
vy2)1/2, but that is not the right expression for |v| in a non-Cartesian space. For example, as will be shown
below in (5.10.5), |eθ'| = |eθ| , so the magnitude of the vector e'θ shown on the left above is in fact |eθ'| = r
= 2 and not 1, but the graphical length of the arrow is 1 since e'θ = (1,0). See Section C.5 for further
discussion of this topic with a specific 2D non-orthogonal coordinate system.

(c) In this example, two basis vectors e'n in x'-space on the left map into the two en vectors on the right
according to en ≡ Se'n. If one were to apply the full mapping x = F-1(x') to each point along the arrows
e'n, for some general non-linear F one would find that these arrows map into warped arrows on the right
whose bases are tangent to those of the en. Those warped arrows lie on the coordinate lines. For this
particular mapping, e'θ maps under F-1 into the warped gray arrow, while e'r maps into er.

Example 2: Spherical Coordinates, tangent base vectors

x = (x1, x2, x3 ) = (x,y,z)

x' = (x1', x2',x3') = (r,θ,φ)

x = F-1(x') ↔ x = rsinθcosφ
y = rsinθsinφ
z = rcosθ (1.6)

S11= (∂x/∂r) = sinθcosφ Sik ≡ (∂xi/∂x'k)

S12 = (∂x/∂θ) = rcosθcosφ
S13 = (∂x/∂φ) = -rsinθsinφ
S21= (∂y/∂r) = sinθsinφ
S22 = (∂y/∂θ) = rcosθsinφ
S23 = (∂y/∂φ) = rsinθcosφ
S31= (∂z/∂r) = cosθ
S32 = (∂z/∂θ) = -rsinθ
S33 = (∂z/∂φ) = 0 (3.4.4)

⎛ sinθcosφ rcosθcosφ -rsinθsinφ ⎞ ⎛ sinθcosφ sinθsinφ cosθ ⎞

S = ⎜ sinθsinφ rcosθsinφ rsinθcosφ ⎟ R = ⎜ cosθcosφ/r cosθsinφ/r -sinθ/r ⎟
⎝ cosθ -rsinθ 0 ⎠ ⎝ -sinφ/(rsinθ) cosφ/(rsinθ) 0 ⎠

where Maple computes R as S-1 and finds as well that: det(S) = r2 sinθ .

Again, from (3.2.7) the tangent base vectors are the columns of S, so

35
 Chapter 3: Tangent Base Vectors

er = (sinθcosφ, sinθsinφ,cosθ) |er| = 1 ≡ h'r

eθ = r(cosθcosφ,cosθsinφ,-sinθ) |eθ| = r ≡ h'θ
eφ = rsinθ(-sinφ,cosφ,0) |eφ| = rsinθ ≡ h'φ (3.4.5)

and unit vector versions are then

^
e r = (sinθcosφ, sinθsinφ,cosθ) =^
r er = ^
r
^eθ = (cosθcosφ,cosθsinφ,-sinθ) =^
θ eθ = r ^
θ
^eφ = (-sinφ,cosφ,0) ^
=φ ^
eφ = rsinθ φ (3.4.6)

The unit vectors can be displayed in this standard picture,

(3.4.7)

^, ^
Notice that (r ^) = (e
θ, φ ^ 1, ^
e 2, ^
e 3) form a right-handed coordinate system at the point x = r.

3.5 The inverse tangent base vectors u'n and inverse coordinate lines

A complete swap x' ↔ x for a mapping x' = F(x) of course produces the "inverse mapping". This has the
effect of causing R ↔ S in the above discussion. The tangent base vectors for the inverse mapping would
then be the columns of matrix R instead of S. We shall denote these inverse tangent base vectors which
exist in x'-space by the symbol u'n. Then:

(en)i = Sin = ∂xi/∂x'n // the tangent base vectors as above (3.2.6)

S = [e1, e2, e3 .... eN ] // are the columns of S (3.2.7)

(u'n)i = Rin = ∂x'i/∂xn // inverse tangent base vectors

R = [u'1, u'2, u'3 .... u'N ] // are the columns of R (3.5.1)

By varying only xn in x-space holding all the other xi = constant, one generates the xn-coordinate lines in
x'-space, just the reverse of the earlier discussion of this subject. Then inverse tangent base vectors u'n
will then be tangent to these inverse coordinate lines. An example is given just below and another appears
in Appendix C.

36
 Chapter 3: Tangent Base Vectors

The vector en was shown to transform as a contravariant vector into an axis-aligned basis vector e'n in x'-
space
(3.3.2) (3.3.3) (3.2.6) (3.2.1)
e'n = R en (e'n)i = ΣjRij (en)j (en)i = Sin (e'n)i = δn,i (3.5.2)

The same thing happens here, only in reverse :

un = S u'n (un)i = ΣjSij (u'n)j (u'n)i = Rin (un)i = δn,i (3.5.3)

where now the un are axis-aligned basis vectors in x-space. A prime on an object indicates which space it
inhabits.

The inverse tangent base vectors u'n are not the same as the reciprocal base vectors En introduced in
Chapter 6 below.

Example 1: Polar coordinates: inverse tangent base vectors and inverse coordinate lines

It was shown earlier for polar coordinates that

-sinθ/r cosθ/r
R = S-1 = ⎛⎝ cos(θ) sinθ ⎞⎠ (3.4.1)

so the inverse tangent base vectors (expressed here as row vectors as usual to save space) are given by the
columns of R as per (3.5.1),

u'x = ( -sinθ/r,cosθ) // note near θ = 0 that u'x indicates a large negative slope
u'y = (cosθ/r,sinθ) . // note near θ = 0 that u'y indicates a small positive slope (3.5.4)

One expects u'x to be tangent to an inverse coordinate line in x'-space which maps to a line in x-space
along which only x is varying, which is a horizontal line at fixed y (red below). Looking at the small θ
region of the left graph in Fig (3.5.5) below, one sees slopes as just described above.

For the polar coordinates mapping discussed near Fig (3.4.3), horizontal (red) and vertical (blue) lines in
x'-space mapped into circles (red) and rays (blue) in x-space, and the tangent base vectors in x-space were
tangent to the coordinate lines there. If one instead takes horizontal (red) and vertical (blue) lines in x-
space and maps them back into coordinate lines in x'-space, the picture is a bit more complicated. Since y
= rsinθ, the plot of an x-coordinate line (x is varying, y fixed at yi) in x'-space has the form r = yi/sinθ,
where yi denotes some selected y value (a red horizontal line), so plotting r = yi/sinθ in x'-space for
various values of yi displays a set of inverse x-coordinate lines (red). Similarly r = xi/cosθ gives some y-
coordinate lines (blue). Here is a Maple plot:

37
 Chapter 3: Tangent Base Vectors

x'-space (θ,r) x-space (x,y)

(3.5.5)
Another example is given in Appendix C.

38
 Chapter 4: Cartesian Space

4. Notions of length, distance and scalar product in Cartesian Space

This Section can be interpreted in either Picture B or Picture D where the x-space is Cartesian, G=1.

Up to this point, we have dealt only with the vector space RN (a vector space is sometimes called a linear
space), and have not "endowed" it with a norm, metric or a scalar product. Quantities like dxi above were
just little vectors and x + dx was vector addition.
Now, for the first time (officially), we discuss length and distance, such as they are in a Cartesian
Space, as defined in Chapter 1.
For RN one first defines a norm which determines the "length" of a vector, the first notion of distance
in a limited sense. The "usual" norm is the L2 norm given by

norm of x = || x || ≡ ( x12 + x22 + .... + xN2 )1/2 ≡ |x| . (4.1)

Now we have a normed linear space.

One next defines the notion of the distance between two vectors. Although this can be done in many
ways, just as there are many possible norms, for RN the "natural metric" is defined in terms of the above
L2 norm, so that

distance between x and y = metric = d(x,y) ≡ || x - y ||

= ( [x1-y1]2 + [x2-y2]2 + .... + [xN-yN]2 )1/2 . (4.2)

Now our space is both a normed linear space and a metric space, a combo known as a Banach Space.
One finally adds the notion of a scalar product (inner product) in this way

(x,y) ≡ Σixiyi ≡ x • y // = Σij δi,j xi yj (4.3)

which of course implies this special case,

(x,x) = x • x = Σixi2 = ||x||2 = | x |2 . (4.4)

Our space has now ascended to the higher level of being a real Hilbert Space of N dimensions. All this
structure is implied by the notation RN, our "Cartesian Space".
The length of the vector dx in RN is given by

length of dx = distance between vectors x+dx and x ≡ ds ≡ || dx || = Σi(dxi)2 . (4.5)

To avoid dealing with the square root, one usually writes

(ds)2 ≡ || dx ||2 = Σi(dxi)2 = (dx1)2 + (dx2)2 + ... + (dxN)2

= Σi dxi dxi = Σij δi,j dxi dxj . (4.6)

As shown in the next Section, one can interpret δi,j as the metric tensor in Cartesian Space.

39
 Chapter 4: Cartesian Space

The cursory discussion of this Section is fleshed out in Chapter 2 of Stakgold where the concepts of linear
spaces, norms, metrics and inner products are defined with precision. Stakgold compares our N
dimensional Cartesian Hilbert Space to the N=∞ dimensional Hilbert Spaces used in functional analysis,
where basis vectors might be Legendre polynomials Pn(z) on (-1,1), n = 0,1,2...∞. He has little to say,
however, about curvilinear coordinate spaces in this particular book.

40
 Chapter 5: The Metric Tensor

5. The Metric Tensor

5.1 The Picture D Context

The metric tensor is the heart of the machine of tensor analysis and we shall have a lot to say about it in
this Chapter. Each Section is best presented in the context of one of our Pictures, and there will be some
jumping around between pictures. We apologize for this inconvenience and ask forbearance. Hopefully
the Sections below will give the reader some experience with typical nitty-gritty manipulations. One
advantage of the developmental notation over the standard notation is that matrix methods are easy to use,
and they will be used below. Unless otherwise specified, repeated indices are implicitly summed over
(Einstein convention). But sometimes we do show sum symbols Σ where extra clarity is needed.

We now go to the Picture D context. Comparison with Picture B shows that primes must be placed on
objects F, R and S related to the transformation from x-space to x'-space:

(5.1.1)

The various partial derivatives are determined from their definitions,

R'ik ≡ (∂x'i/∂xk) R"ik ≡ (∂x"i/∂xk) Rik ≡ (∂x"i/∂x'k)

S'ik ≡ (∂xi/∂x'k) S"ik ≡ (∂xi/∂x"k) Sik ≡ (∂x'i/∂x"k) . (5.1.2)

The unprimed S,R can be expressed in terms of the primed objects this way (chain rule, implied Σ on a),

Rik ≡ (∂x"i/∂x'k) = (∂x"i/∂xa) (∂xa/∂x'k) = R"ia S'ak ⇒ R = R" S'

Sik ≡ (∂x'i/∂x"k) = (∂x'i/∂xa) (∂xa/∂x"k) = R'ia S"ak ⇒ S = R' S" . (5.1.3)

5.2 Definition of the metric tensor

The metric or distance between vectors x and x+dx can be specified as done in Chapter 4 in terms of the
norm of differential vector dx,

metric(x+dx, x) = norm( [x+dx] - x) = norm(dx) ≡ | dx | ≡ ds , (5.2.1)

41
 Chapter 5: The Metric Tensor

with the caveat that for non-Cartesian spaces this may not be an official norm, see Section 5.10 below.
The squared distance (ds)2 must be a linear combination of products dxidxj just on dimensional grounds.
The coefficients in this linear combination form a matrix called the metric tensor (later we show this
matrix really is a tensor). Here for an N-dimensional space,

(ds)2 = Σi=1N Σj=1N [ metric tensor ]ij dxi dxj . (5.2.2)

One "endows" a space with a certain metric tensor, and this in turn determines the distance ds = norm(dx)
for any differential vector dx in that space. A metric tensor is specific to a space; it is a property of the
space; it is part of the space's definition.
Recall from the text near (1.10) that our special "quasi-Cartesian" space has a diagonal metric tensor
G whose diagonal elements are independently either +1 or -1. Then distance ds is determined by

(ds)2 = ΣiΣj[Gij]dxidxj = ΣiΣj[Giiδi,j]dxidxj = Σi Gii dxidxi = Σi Gii (dxi)2. (5.2.3)

How might one express this same ds in terms of the the coordinates x'i of x'-space in Picture D? We
found in (2.1.6) that, for Picture A of Fig (2.1.1), dx = S(x')dx' where S is a matrix which represents a
general transformation x' = F(x) locally at a point. Translating this to the right side of Picture D, we have
dx = S'(x') dx' since S is primed there (as is F and R). Therefore

(ds)2 = ΣiGiidxidxi = Σi Gii (ΣkS'ik dx'k) (ΣmS'im dx'm)

= ΣkΣm { Σi Gii S'ikS'im } dx'k dx'm . (5.2.4)

Now let's start all over again and define a metric tensor in x'-space just as we did in x-space,

(ds')2 = ΣkΣm [ metric tensor ]'km dx'k dx'm .

= ΣkΣm ḡ'km dx'k dx'm . (5.2.5)

We call the metric tensor in x'-space g' because x' is primed, and we use an overbar because it will be
shown below that ḡ'km are the components of a covariant rank-2 tensor ḡ' in x'-space.

We now make an important hypothesis: the length of vector dx is the same as the length of the vector dx',
which is to say, we assume that ds = ds'. We are assuming that ds is a scalar under transformation F', so

(ds)2 = (ds')2 . // hypothesis, (ds)2 is a scalar (5.2.6)

Since equations (5.2.4) and (5.2.5) must be valid for any choice of dx', comparison of the two equations
yields this relationship between the two metric tensors ḡ' and G :

ḡ'km ≡ { ΣiGiiS'ikS'im } . (5.2.7)

Using Gij = Giiδi,j we obtain the matrix relationship between metric tensors ḡ' and G :

42
 Chapter 5: The Metric Tensor

ḡ'km ≡ { ΣiGiiS'ikS'im } = ΣiΣj S'TkiGijS'jm ⇒

ḡ' = S'TG S' . (5.2.8)

We now repeat all the above steps for the transformation from x-space to x"-space on the left side of
Picture D in Fig (5.1.1). We make the hypothesis that (ds)2 = (ds")2 and in doing so, we obtain a
relationship between the metric tensor G for x-space and the metric tensor ḡ" for x"-space:

ḡ'' = S''TG S'' . (5.2.9)

To summarize, there are three metric tensors for the three spaces in Picture D :

ḡ = G ḡ' = S'T G S' ḡ" = S"T G S" . (5.2.10)

where we have made the hypothesis that ds = ds' = ds".

Concerning the invariance of (ds). In the above discussion, it was assumed in (5.2.6) that distance (ds)2 is
the same in x'-space as it is in x-space. As shown below in Section 5.10, this is part of a larger hypothesis
that the covariant dot product of any two vectors gives the same number regardless of which space is used
to compute the dot product: A • B = A' • B'. This in turn implies that |A| = |A'| and in particular |dx| =
|dx'| or ds = ds'.
In our major application, where x-space is Cartesian and x'-space is that of some curvilinear
coordinates, it is a requirement that | A | = | A' |. The length of a vector in Cartesian physical space does
not change simply because we choose to express that length in some curvilinear coordinates. Imagine that
A is a velocity vector v. The speed |v| of an object is the same number whether one represents v in
Cartesian or spherical coordinates.
In special relativity it is a very well-verified hypothesis that dot products are scalars and that (ds)2 is a
scalar under Lorentz transformations. In this context, ds is often written dτ (the so-called proper time).
There are, however, applications of transformations where the scalarity of (ds)2 is not valid and in
fact it is crucial that (ds)2 can change under a transformation. For example, in continuum mechanics one
can think of the flow of a tiny bit of continuous matter as being modeled by a transformation x = F(x',t).
In general F ( = F-1) is non-linear. At time t = 0 the geometry of this little blob is described by x'-space
coordinates, and after a flow at time t = t it is described by x-space coordinates. If once traces during this
flow a little "dumbbell" vector between two very closely spaced particles in the blob, one finds that dx' at
time t = 0 becomes dx = Sdx' at time t = t where S is not simply a rotation. The whole point here is that
during the flow, the distance vector between two close particles rotates and stretches in some manner, and
in general (due to this stretch) |dx| ≠ |dx'|, so the geometric (ds)2 is definitely not invariant under the flow
(ie, under the transformation F). On the other hand, as we shall see below, a certain curvilinear (ds)2 does
remain invariant under the flow. The flow application is considered more in Section 5.16 below, and still
further in Appendix K.

43
 Chapter 5: The Metric Tensor

5.3 Inverse of the metric tensor

The inverses of the three metric tensors in (5.2.10) shall be indicates without an overbar, and we shall
eventually show these matrices to be "contravariant" matrices and thus deserve no overbar. We thus now
define three new g matrices as these inverses, and compute the inverses:

g ≡ ḡ -1 = G-1 = G // remember G just has +1 and -1 diagonal elements

g' ≡ ḡ'-1 = (S'T G S')-1 = S'-1 G (S'T)-1 = R' G R'T

g" ≡ ḡ"-1 = (S"T G S")-1 = S"-1 G (S"T)-1 = R" G R"T . (5.3.1)

Here are the collected facts from above:

g=G g' = R'G R'T g" = R" G R"T S = R' S"

ḡ = G ḡ' = S'TG S' ḡ" = S"T G S" R = R" S'
ḡg = 1 ḡ'g' = 1 ḡ"g = 1 (5.3.2)

Comment: In the Picture C context but with a Quasi-Cartesian x(0)-space (g(0)=G), one could take the
second column above and write it this way,

g = RGRT
ḡ = STGS
ḡg = 1 (ds)2 = ḡ km dxk dxm (5.3.3)

where now the clutter of primes is gone. If x(0)-space is Cartesian so g(0) = G = 1, then

g = RRT
ḡ = STS. // if x(0)-space is Cartesian in Picture C (5.3.4)

But we continue with Picture D shown in (5.1.1).

5.4 A metric tensor is symmetric

Consider the following, where A is any matrix and D is a diagonal matrix :

N = ADAT ⇒ NT = (ADAT)T = ADTAT = ADAT = N ⇒ N symmetric (5.4.1)

Replacing A with AT and N with M gives

44
 Chapter 5: The Metric Tensor

M = ATDA ⇒ MT = (ATDA)T = ATDT(AT)T = ATDA = M ⇒ M symmetric (5.4.2)

Looking at (5.3.3), since g has the form of N and ḡ the form of M with G diagonal, we conclude that both
g and ḡ are symmetric matrices, and the same is of course true for g' and g". Thus

gab = gba ḡ ab = ḡ ba
g'ab = g'ba ḡ'ab = ḡ'ba (5.4.3)

5.5 det(g) and gnn of a Cartesian-generated metric tensor are non-negative

If we arrive at x'-space by a transformation F from a Cartesian x-space (as opposed to a Quasi-Cartesian

one), we refer to the metric tensor g' in this x'-space as being "Cartesian generated". In this case G = 1 and
the metric tensors above are g = RRT and ḡ = STS as in (5.3.4). Any matrix of either of these forms has
positive diagonal elements and positive determinant:

(ATA)aa = Σb (AT)abAba = Σb (A)baAba = Σb (Aba)2 ≥ 0 // diagonal elements ≥ 0

det(ATA) = det(AT) det(A) = det(A) det(A) = [ det(A) ]2 ≥ 0 // det ≥ 0 . (5.5.1)

To show these results for the AAT form, just replace A→AT everywhere. Recall that transformation F
maps RN → RN so the coefficients of the linearized matrices R and S are real, and elements of the metric
tensor must therefore also be real. For a Quasi-Cartesian-generated metric tensor, these proofs are invalid
since then g = RGRT and ḡ = STGS and G ≠1.

5.6 Definition of two kinds of rank-2 tensors

We now switch to Picture A,

(5.6.1)

Recall the vector transformation rules from (2.5.1),

V' = R V contravariant Rik(x) ≡ (∂x'i/∂xk) R = S-1

V̄' = ST V̄ covariant Sik(x') ≡ (∂xi/∂x'k) = STki(x') (2.5.1)

which can be written out in components (implied sum on a'),

45
 Chapter 5: The Metric Tensor

V'a = Raa' Va' contravariant Rik(x) ≡ (∂x'i/∂xk) R = S-1

V̄'a = STaa'V̄a' covariant Sik(x') ≡ (∂xi/∂x'k) = STki(x') . (5.6.2)

A rank-1 tensor is defined to be a vector which transforms in one of the two ways shown above.
Similarly, a (non-mixed) rank-2 tensor is defined as a matrix which transforms in one of these two ways:

M'ab = Raa' Rbb' Ma'b' // contravariant rank-2 tensor

M̄'ab = STaa' STbb' M̄a'b' // covariant rank-2 tensor (5.6.3)

and again we put a bar over the covariant objects.

Digression: Proof that (A-1)T = (AT)-1 for any invertible matrix A:

(5.6.4)
• det(A) = det(AT)
• cof(AT) = [ cof(A)]T since [cof(AT)]ab = cof ( ATab) = cof(Aba) = [cof(A)]ba = [cof(A)]Tab
• (A-1)T = { [cof(A)]T / det(A) }T = [cof(AT)]T /det(AT) = (AT)-1

This fact is used many times in the manipulations below.

5.7 Proof that the metric tensor and its inverse are both rank-2 tensors

The above rank-2 tensor transformation rules (5.6.3) can be written in the following matrix form
(something not possible with higher-rank tensors),

M' = R M RT // contravariant rank-2 tensor

M̄' = ST M̄ S // covariant rank-2 tensor (5.7.1)

where recall

. (5.7.2)

But we now switch these rules to the Picture D context (upper arrow) where F maps x'-space to x"-space,

46
 Chapter 5: The Metric Tensor

(5.1.1)
to obtain

M" = R M' RT // contravariant rank-2 tensor (5.7.3)

M̄" = ST M̄' S . // covariant rank-2 tensor

Consider then this sequence of steps:

1 *G * 1 = 1 * G * 1
(S"R") G (S"R")T = (S'R') G (S'R')T // S"R" = 1 and S'R' = 1
S" (R"G R"T) S"T = S'(R'G R'T) S'T // regroup
S" g" S"T = S' g' S'T // since g" = R"G R"T and g' = R'G R'T by (5.3.2)
g" S"T = R" S' g' S'T // left multiply by S"-1 = R"
g" = R" S' g' S'T R"T // right multiply by S"T,-1 = R"T (5.7.4)
g" = (R" S') g' (S'T R"T) // regroup
g" = (R" S') g' (R" S')T // (AB)T = BTAT
g" = R g' RT // using long forgotten (5.1.3)

With definition (5.7.3) this last result then shows that g' is a contravariant rank-2 tensor with respect to
the transformation F taking x'-space to x"-space. Continuing on,

g" = R g' RT
g"-1 = (R g' RT)-1
g"-1 = ST g'-1 S // RT,-1= ST etc
ḡ" = ST ḡ' S // ḡ' = g'-1 (5.7.5)

and this last result shows that ḡ' is a covariant rank-2 tensor with respect to the transformation F taking
x'-space to x"-space, again from (5.7.3). This is why we put a bar over this g from the start.

These two metric tensor transformation statements can be converted to the Picture A context,

47
 Chapter 5: The Metric Tensor

g' = R g RT g'ab = Raa'Rbb'ga'b' // g is a contravariant rank-2 tensor

ḡ' = ST ḡ S ḡ'ab = STaa'STbb' ḡ a'b' // ḡ is a covariant rank-2 tensor (5.7.6)

Since RS = 1, the equations can be inverted to get

g = S g' ST gab = Saa'Sbb'g'a'b'

ḡ = RT ḡ' R ḡ ab = RTaa'RTbb' ḡ'a'b' = Ra'aRb'b ḡ'a'b' (5.7.7)

Further variations of the above are obtained using RS = 1, gḡ = g'ḡ' = 1 and g = gT (etc.) :

Rg = g' ST ḡ' R g = ST g' ST ḡ = R

ḡ S = RT ḡ' g RT ḡ' = S ḡ S g' = RT . (5.7.8)

Finally, if x-space is Cartesian so g = ḡ = 1, one has, analogous to (5.3.4) for Picture C,

g' = RRT // g = 1
ḡ' = STS (5.7.9)

Comment: If x-space and x'-space are both quasi-Cartesian, so g = g' = G, then (5.7.6) says G = RGRT.
This does not imply that R must be independent of x. One could have G = R(x)GR(x)T. In particular, if
both spaces are Cartesian so g = g' = 1, then one can have 1 = R(x)R(x)T if R(x) is a real orthogonal
matrix for every value of x, R-1(x) = RT(x). When R-1 = RT we refer to R as a "rotation", where we
include possible parity transformations. Then if g = 1 and g' = 1 one can have x' = F(x) = R(x) x where we
refer to R(x) as a local rotation. The rotation can be different at every point in space. An example would
be

⎛ cosθ -sinθ 0 ⎞
R(x) = Rz(θ) = ⎜ sinθ cosθ 0 ⎟ where θ = θ(x) . (5.7.10)
⎝ 0 0 1 ⎠

For N = 4 dimensions in special relativity where g = g' = G = diag(-1,1,1,1), one can have 4x4 boost and
rotation matrices of the form shown below in (5.14.3) and (5.14.4) where the parameters b and r could be
b(x) and r(x).

48
 Chapter 5: The Metric Tensor

5.8 Metric tensor converts vector types

We continue in Picture A. Suppose V is a contravariant vector so V' = RV. Construct a new vector W
with the following properties

W = ḡ V x-space
W' = ḡ' V' x'-space . (5.8.1)

Is vector W one of our two vector types, or is it neither? One must examine how it transforms under F:

W' = ḡ' V' = (ST ḡ S) (RV) = ST ḡ (SR)V = ST ḡ V = ST W . // using (5.7.6) (5.8.2)

Therefore from (2.5.1) this new vector W is a covariant vector under F, so it should have an overbar,

W̄ ≡ ḡ V and W̄' = ST W̄ . (5.8.3)

This covariant vector W̄ can be regarded as the covariant partner of contravariant vector V.

This shows the general idea that applying ḡ to any contravariant vector produces a covariant vector! So
this is one way to construct covariant vectors if we have a supply of contravariant ones. Conversely,
starting with a known covariant vector W̄, one can construct a contravariant vector V ≡ g W̄ . Thus,
every vector of either type can be thought of as having a partner vector of the other type.

An obvious notation is to write W̄ as V̄ so no extra letter is needed. Then one has

V̄ = ḡ V V = g V̄ // ḡ = g-1 from (5.3.1)

V̄i = ḡ ijVj Vi= gijV̄j . (5.8.4)

Then the vectors V and V̄ are the partner vectors, the first contravariant, the second covariant.

5.9 Vectors in Cartesian space

Theorem: There is no distinction between a contravariant and a covariant vector in Cartesian space.
(5.9.1)
Proof: Pick a contravariant vector V. Since ḡ = 1, V̄ ≡ ḡ V = V . But V̄ is a covariant vector. Since V̄ =
V, every contravariant vector is identical to its covariant partner in x-space. The vector components are
numerically equal. The transformation rules (2.5.1) in this case are

V' = R V
V̄' = ST V̄ = ST V . (5.9.2)

Since in general one does not have R = ST, one sees that in general V' ≠ V̄' , so the two partner vectors are
in general not identical in x'-space even though they are identical in Cartesian x-space. In the special case
that R is a rotation, R = R-1,T (real orthogonal) and then R = R-1,T = ST so V' = V̄' .

49
 Chapter 5: The Metric Tensor

5.10 The covariant dot product A • B and norm |A|

For a Cartesian space, Chapter 4 defined the norm as the length of a vector, the metric as the distance
between two vectors, and the scalar product (inner product) as the projection of one vector on another.
The official definitions of norm, metric and scalar product require non-negativity: |x| ≥ 0, d(x,y) ≥ 0, and
x • x ≥ 0. For non-Cartesian spaces, the logical extensions of these three concepts can result in all three
quantities being negative. Nevertheless, we shall use the term "covariant scalar product" with notation A •
B as defined below, as well as the notation |A|2 ≡ A • A where |A| will be called the length, magnitude
or norm of A, even though these objects are not true scalar products or norms. In the curvilinear
application of tensor analysis, where x-space is Cartesian, since the norm and scalar product are tensorial
scalars, and since they are non-negative in Cartesian x-space, the problem of negative norms does not
arise in either space.
How do authors handle this problem? Some authors refer to A • A as "the norm" of A (e.g., Messiah
last line of p 878 discussing special relativity), which is our |A|2. For a general 4-vector A in special or
general relativity, most authors just write A • A (AμAμ in standard notation), they note that the quantity is
invariant under transformations, but don't give it a name.
Whereas we use the bold • for this covariant dot product, most special relativity authors prefer to
reserve this bold dot for a 3D spatial dot product, and then the 4D dot product is written with some "less
bold dot" such as A.B or A•B. Typical usage then in standard notation would be p•p = pμpμ = p02 - p•p
(see for example Bjorken and Drell p 281).

Without further ado, we define the "covariant scalar product" of two contravariant vectors (a new and
different use of the word "covariant", but the same as appears in Section 7.15) as

A • B ≡ AaB̄a . // implied sum on a (5.10.1)

The important fact about the dot product of two tensorial vectors is that it is a tensorial scalar, as we now
show using (5.9.2) for the transformations of A and B and (2.1.6) that SR = 1 :

A' • B' ≡ A'aB̄a' = (RA)a(STB̄)a = (RaiAi)(SjaB̄j) = (SjaRai)AiB̄j = (SR)jiAiB̄j

= δj,iAiB̄j = AiB̄i = A • B . (5.10.2)

There are various equivalent ways to write the dot product of (5.10.1) using (5.8.4) that V̄ = ḡ V and
conversely that V = g V̄ , and also the fact (5.4.3) that gab = gba :

50
 Chapter 5: The Metric Tensor

A • B ≡ AaB̄a = Aa(ḡ abBb) = ḡ abAaBb = (ḡ baAa)Bb = ĀbBb = ĀaBa

A • B ≡ AaB̄a = (gĀ)aB̄a = gab Āb B̄a = gab Āb B̄a .

To summarize

A • B = AaB̄a = ĀaBa = ḡ abAaBb = gab Āb B̄a . (5.10.3)

In the special case that A = B, we use the shorthand norm notation (with caveat as noted above) and
(5.10.2) to obtain,

|A|2 ≡ A • A = A' • A' = |A'|2 . (5.10.4)

Going back to Chapter 3 and the vectors e'n and en, a claim made at the start of Section 3.4 can now be
verified:

|e'n|2 = e'n • e'n = en • en = |en|2 ⇒ |e'n| = |en| . (5.10.5)

Comment on Notation

Consider again the definition (5.10.1) and the alternate form ĀaBa shown in (5.10.3)

A • B ≡ AaB̄a = ĀaBa . (5.10.3)

The dot product involves the contravariant components of one vector and the covariant components of the
other vector. In the dot product notation A • B , we have indicated each vector by its contravariant name
just as a convention. We could just as well have indicated one or both vectors by its covariant name, but
the dot product indicated by whatever name would be the same: contravariant components of one vector
and the covariant components of the other vector. Thus,

A • B ≡ AiB̄i = ĀiBi = Ā • B = A • B̄ = Ā • B̄ . (5.10.6)

We shall always use the first notation A • B since it is the simplest. Similarly, when B = A,

|A|2 ≡ A • A = |Ā|2 . (5.10.7)

We thus have the interesting fact that, if A and B are tensorial vectors for general x' = F(x), then

A ≠ Ā but A • B = Ā • B for any B . (5.10.8)

If B = A, this says

A ≠ Ā but A • A = Ā • A = |A|2 = AiĀi = gijAiAj . (5.10.9)

51
 Chapter 5: The Metric Tensor

We mention this notational issue to head off the following incorrect notion:

A • B = ḡ abAaBb ⇒ Ā • B̄ = ḡ abĀaB̄b ≠ A • B . // wrong!!!

A dot product application

In applications in which (ds)2 is regarded as a scalar with respect to transformation F we have

(ds')2 = dx' • dx' = (ds)2 = dx • dx (5.10.10)

and ds = ds' is called "the invariant distance". Such applications include curvilinear coordinate
transformations and relativity transformations.

In special relativity, using the Bjorken and Drell notation noted above where g'μν = diag(1,-1,-,1,-1) and
c = 1, one writes (Standard Notation),

(dτ)2 = g'μν dx'μdx'ν = dx'μdx'μ = dx'• dx' = dx • dx = a Lorentz scalar = (dt)2 - dx • dx , xμ = (t,x)
(5.10.11)

and dτ is called "the proper time", a particular case of the invariant distance ds. Notice that (dτ)2 < 0 for a
spacelike 4-vector dxμ, meaning one that lies outside the future and past lightcones (|dx| > |dt| ). [We now
restore • to our covariant definition after temporarily using it above for a 3-space Cartesian dot product. ]

5.11 Metric tensor and tangent base vectors: scale factors and orthogonal coordinates

The context of Picture A continues,

(5.11.1)

Recall this fact from (3.2.7),

S = [e1, e2, e3 .... eN ] (3.2.7)

where the columns of S are the tangent base vectors. It follows from (5.7.6) that

ḡ' = ST ḡ S = [e1, e2, e3 .... eN ]T ḡ [e1, e2, e3 .... eN ] (5.11.2)

so

52
 Chapter 5: The Metric Tensor

e1•e1 e1• e2 e1 • e3 ...... e1• eN

e2•e1 e2• e2 e2 • e3 ...... e2• eN
ḡ' = e3•e1 e3• e2 e3 • e3 ...... e3• eN (5.11.3)
........
eN•e1 eN• e2 eN • e3 ...... eN• eN

since,

enT ḡ em = (en)i ḡ ij (em)j = ḡ ij(en)i(em)j = en • em (5.11.4)

using the covariant dot product shown in (5.10.3). Taking the m,n component of (5.11.3) one gets

ḡ'mn = em • en or ḡ'mn = ∂'mx • ∂'nx // using (3.2.6) (5.11.5)

which makes a direct connection between the covariant metric tensor in x'-space and the tangent base
vectors en in x-space. A less graphical derivation of this fact uses (5.7.6) and (3.2.5),

ḡ'nm = (ST ḡS)nm = STna ḡ ab Sbm = ḡ ab San Sbm = ḡ ab (en)a(eb)n ≡ en • em . (5.11.6)

Scale Factors and Orthgonal Coordinates

We showed just above in (5.11.5) that em • en = ḡ'mn. Regardless of whether or not the coordinates are
orthogonal, one can define the scale factor h'n and unit vector ^
e n as follows,

h'n2 ≡ |en|2 = en • en = ḡ'nn

⇒ h'n = |en| = ḡ'nn
⇒ ^ ^ n| = 1 .
e n ≡ en/h'n = "unit vector", |e (5.11.7)

If the tangent base vectors en are orthogonal, then clearly the metric tensor ḡ'mn = em • en must be
diagonal. In fact, based on (5.11.7), the covariant metric tensor must be

ḡ'mn = h'm2 δm,n .

We claim that the contravariant metric tensor g'nm is then also diagonal with diagonal elements

g'nn = h'n-2 .

To verify this claim, we confirm that ḡ' g' = 1:

Σm ḡ'nm g'mk = Σm [δn,m h'm2][δm,kh'm-2] = Σm δn,mδm,k = δn,k = (1)nk

Since the inverse matrix is unique as long as det(g')≠0, g'mk = δm,kh'm-2 must be it!

53
 Chapter 5: The Metric Tensor

We have then established this simple fact:

orthogonal coordinates ⇔ ḡ'mn = h'm2 δm,n covariant

g'mn = h'm-2δm,n contravariant . (5.11.8)

For an orthogonal coordinate system, at any point x in x-space, the tangents en to the N coordinate lines
passing through that point are orthogonal. Most examples below will involve such systems, with
Appendix C providing a non-orthogonal example.

In the Standard Notation of Chapter 7, the above association becomes

orthogonal coordinates ⇔ g'mn = h'm2 δm,n covariant

g'mn = h'm-2δm,n contravariant . (5.11.9)

The reason we put a prime on h'n is because it is associated with x'-space and its metric tensor g' in
Picture A or B of Fig (1.11). The curvilinear coordinates are x'.
When working with Picture C, however, we would call this scale factor hn because it is then
associated with the metric tensor g and x-space. In Picture C, the curvilinear coordinates are x.

In other Pictures, the curvilinear "space on the left" might be called ξ-space, and since this does not carry
a prime, again one would write hn.

Section 5.5 above showed that ḡ'nn ≥ 0 when x-space is Cartesian. This is the usual case for the
curvilinear coordinates application, and so in this case the scale factors h'n are always real and positive.

Note 1. Different authors use different symbols for scale factors. For example, Margenau and Murphy
refer to them as the Qn. Morse and Feshbach and most modern works use hn. They never use a prime,
because they always use Picture C or equivalent where x is the curvilinear coordinate.

Note 2: Some authors refer to the scale factors h'n as the Lamé coefficients, while other authors refer to
Rij as the Lamé coefficients which they call hji. (Lame, for PDF search with no accent mark)

54
 Chapter 5: The Metric Tensor

5.12 The Jacobian J

The context of Picture A continues,

(5.12.1)

First of all, note that since RS = 1,

det(S) = 1/det(R) . (5.12.2)

Recall from (2.1.6) our definitions of Rik and Sik :

Rik(x) = (∂x'i/∂xk) is called the Jacobian matrix for the transformation x' = F(x)
Sik(x') = (∂xi/∂x'k) is then the Jacobian matrix of the inverse transformation x = F-1(x') . (5.12.3)

There are two Jacobian matrices here, which are inverses of each other since RS = 1, and for each of these
matrices we could define "a Jacobian" as the determinant of that matrix:

J(F) = det(R) 1/ J(F) = 1/det(R) = det(S) = J(F-1)

J(F-1) = det(S) . (5.12.4)

Because our major interest is in curvilinear coordinates x', and because curvilinear coordinates are almost
always defined by equations of the inverse form x = F-1(x') such as the following for polar coordinates,

x = rcosθ x = F-1(x') (5.12.5)

y = rsinθ,

we shall officially define "the Jacobian" to be J = J(F-1). So,

the Jacobian ≡ J(x') ≡ det(S(x')) = det(∂xi/∂x'k) = 1/det(R(x(x')) = 1/ det(∂x'i/∂xk) . (5.12.6)

Note 1: Objects which relate to the transformation between x-space and x'-space cannot themselves be
tensors because tensor objects must be associated with a specific space, the way V(x) is a vector in x-
space and V'(x') is a vector in x'-space. Thus Sij(x') = ∂xi/∂x'k , although a matrix, is not a rank-2 tensor.
Similarly, J(x'), while a "scalar" function, is not a rank-0 tensorial scalar. One does not ask how S and J
themselves "transform" in going from x-space to x'-space.

Note 2: An alternative notation used by some authors is this

J(x,x') ≡ det(S(x,x')) = det(∂xi/∂x'k) (5.12.7)

55
 Chapter 5: The Metric Tensor

as if x and x' were independent variables. In our presentation, x is not an independent variable but is
determined by x = F-1(x'). Just as one might write f '(x') = ∂f/∂x', we write J(x') = det(∂xi/∂x'k). The
connection would be J(x') = J(x=F-1(x'),x')) = J(x(x'),x').

Note 3: Other sources often use the notation | M | to indicate the determinant of a matrix. We shall use the
notation det(M), and reserve | | to indicate the magnitude of some quantity, such as |J| below.

The determinant of any NxN matrix S may be written (εabc.. is the permutation tensor, see Section 7.7),

det(S) = εabc...x Sa1 Sb2 ... SxN . // = εabc...x S1a S2b ... SNx (5.12.8)

For our particular S with Sin = (en)i from (3.2.5) this becomes

det(S) = εabc...x (e1)a(e2)b....... (eN)x (5.12.9)

so J is related to the tangent base vectors by

J = εabc...x (e1)a(e2)b....... (eN)x . (5.12.10)

It was shown in (5.7.6) that ḡ' = ST ḡ S and g' = R g RT , these being the transformation rules for
covariant and contravariant rank-2-tensors. Therefore

det(ḡ') = det(ST ḡS) = det(ST)det(ḡ)det(S) = det(S)det(S)det(ḡ) = J2 det(ḡ)

det(g') = det(RgRT) = det(R)det(g)det(RT) = det(R)det(R)det(g) = J-2 det(g)
or
det(ḡ') = J2 det(ḡ) ⇒ J2 = det(ḡ') / det(ḡ) = [det(S)]2
det(g') = J-2det(g) . (5.12.11)

It is a tradition to define certain scalar (but not tensorial scalar) objects with the same name g and g',

g(x) ≡ det(ḡ(x)) = 1/det(g(x)) // in x-space

g'(x') ≡ det(ḡ'(x')) = 1/det(g'(x')) // in x'-space (5.12.12)

so that

J2(x') = det(ḡ'(x')) / det(ḡ(x)) = g'(x') / g(x) . (5.12.13)

Normally the argument dependence is suppressed and one then writes

g ≡ det(ḡ) = 1/det(g)
g' ≡ det(ḡ') = 1/det(g')
J2 = det(ḡ')/ det(ḡ) = g'/g ⇒ g' = J2 g . (5.12.14)

As explained in Section D.1, the equation g' = J2 g says that g, instead of being a tensorial scalar, is a
scalar density of weight -2. A tensorial scalar s has weight 0: s' = J0s = s.

56
 Chapter 5: The Metric Tensor

Warning: One must be careful to distinguish the scalars g and g' from the the matrices g and g'. It is
usually clear from the context which meaning is implied.

It is convenient to make the following definition, called the signature of the metric tensor,

s = sign[det(ḡ)] = sign(g) // that is, s = either +1 or -1 . (5.12.15)

Since g ḡ = 1 by (5.3.2), one has det(g)det(ḡ) = 1 or so that sign[det(ḡ)] = sign[det(g)] .

Since det(ḡ') / det(ḡ) = [det(S)]2 by (5.12.11), one has sign[det(ḡ)] = sign[det(ḡ')].
Since g' ḡ' = 1 by (5.3.2), one has det(g')det(ḡ') = 1 or so that sign[det(ḡ')] = sign[det(g')] .

Therefore:

s = sign[det(ḡ)] = sign[det(g)] = sign[det(ḡ')] = sign[det(g')] = sign(g) = sign(g') . (5.12.16)

Since transformation F is assumed invertible in its domain and range, one cannot have det(S)=0 anywhere
except perhaps on a boundary. Since det(ḡ') = [det(S)]2det(ḡ) by (5.12.11), if we assume det(ḡ) is non-
vanishing in the x-space domain of F, then det(ḡ') ≠ 0 everywhere in the range of F. The conclusion with
this assumption is that the signature s is always well-defined.

Obviously, the quantities sg and sg' are both positive, and since J2 = g'/g one can write

|J| = sg' / sg = | det(S) | = g'/g . (5.12.17)

For the curvilinear coordinates application, x-space is Cartesian, det(ḡ) = g = 1, and thus s = 1 and then

|J| = g' = | det(S) | . // curvilinear (5.12.18)

For the relativity application, x-space is Minkowski space with det(ḡ) = g = -1 so s = -1 and

|J| = -g' = | det(S) | . // relativity (5.12.19)

Here then is a summary of the results of this Section:

57
 Chapter 5: The Metric Tensor

J(x') ≡ det(S(x')) = det(∂xi/∂x'k) = 1/det(R(x(x')) = 1/ det(∂x'i/∂xk)

g ≡ det(ḡ) g' ≡ det(ḡ')

g' = J2g ⇒ g is a scalar density of weight -2 (Section D.1)

s ≡ sign[det(ḡ)] = sign[det(g)] = sign[det(ḡ')] = sign[det(g')] = sign(g) = sign(g')

|J| = sg' / sg = | det(S) | = g'/g (5.12.20)

Note: Weinberg p 98 (4.4.1) defines g = -det(gij). This is the only one of Weinberg's conventions that we
have not adopted, so in this paper it is always true that g ≡ + det(gij) even though this is -1 in the
application to special relativity.

Carl Gustav Jacob Jacobi (1804 –1851). German, Berlin PhD 1825 then went to Konigsberg, did much in
a short life. Elucidated the whole world of elliptic integrals and functions, such as F(x,k) and sn(x;k),
which occur even in simple problems like the 2D pendulum. Wiki claims he promoted Legendre's ∂
symbol for partial derivatives (used throughout this document) and made it a standard. Among many
other contributions, he saw the significance of the object J which now bears his name: "the Jacobian". The
Jacobi Identity is another familiar item, a rule for non-commuting operators [x,[y,z]] + [z,[x,y]] + [y,[z,x]]
= 0 which finds use with quantum mechanical operators and matrices, and more generally with Lie group
generators -- examples of which appear in Section 5.14 below.

5.13 Some relations between g, R and S in Pictures B and C (Cartesian x-space).

In Picture B, which is Picture A with g = 1,

(5.13.1)

the statement of the rank-2 tensor transformation of g' and ḡ' becomes, as shown in (5.7.9),

g' = RRT
ḡ' = STS (5.13.2)

which can be written in a variety of ways,

RT = (SR)RT = S(RRT) = S g' ⇒ R = g' ST ⇒ 1 = S g' ST

ST = ST(RTST) = (STS)R = ḡ' R ⇒ S = RT ḡ' ⇒ 1 = RT ḡ' R . (5.13.3)

In summary:

58
 Chapter 5: The Metric Tensor

g' = RRT RT = S g' R = g' ST 1 = S g' ST

ḡ' = STS ST = ḡ' R S = RT ḡ' 1 = RT ḡ' R . (5.13.4)

The diagonal elements of ḡ' and g' are given by

ḡ'nn = Σn STniSin = Σn (Sin2) = Σi (∂xi/∂x'n)2

g'nn = Σn RniRTin = Σn (Rni2) = Σi (∂x'n/∂xi)2 . (5.13.5)

If the x'i are orthogonal coordinates, then ḡ'nm = h'n2δn,m and g'nm = h'n-2δn,m as shown in (5.11.8)
where the h'n are the scale factors defined in (5.11.7). These scale factors may then be expressed as,

h'n2 = ḡ'nn = Σi (∂xi/∂x'n)2 h'n-2 = g'nn = Σi (∂x'n/∂xi)2 . (5.13.6)

With notational changes x'n → ξn (curvilinear coordinates for M&F) and h'n → hn, these last two
equations appear (for three dimensions) in Morse & Feshbach Vol I p 24 equation (1.3.4):

(5.13.7)

In Picture A (5.13.1) the curvilinear coordinates are called x'n and the Cartesian coordinates are xn. In
Picture C, the curvilinear coordinates are called xn and the Cartesian coordinates are xn(0). To simplify
notation, we use ξn in place of xn(0) for the Cartesian coordinates (these are totally different ξn from the
ξn shown above in the M&F quote).

ξn ≡ xn(0) (5.13.8)

In this Picture C, equations (5.13.2,4,5,6) appear as follows (no primes on g's) :

g = RRT
ḡ = STS

g = RRT RT = S g R = g ST 1 = S gST
ḡ = STS ST = ḡ R S = RT ḡ 1 = RT ḡ R

ḡ nn = Σn STniSin = Σn (Sin2) = Σi (∂ξi/∂xn)2

gnn = Σn RniRTin = Σn (Rni2) = Σi (∂xn/∂ξi)2

hn2 = ḡ nn = Σi (∂ξi/∂xn)2 hn-2 = gnn = Σi (∂xn/∂ξi)2 (5.13.9)

59
 Chapter 5: The Metric Tensor

In Picture B of (5.13.1), we have primes on things like g' and hn' because these primes match the names
of the curvilinear coordinates which are x'n and which live in x'-space. In Picture C, since the curvilinear
coordinates are now xn, objects like g and hn have no primes to match the fact that xn have no primes and
live in x-space. And in Picture C, one has Sin = ∂ξi/∂xn, where xn is the curvilinear coordinate and ξi the
Cartesian coordinate. It is admittedly a bit confusing, but it is just notation.

Example 1: Polar coordinates: metric tensor and Jacobian

Picture C (5.13.8) continues (so now θ = x1 and r = x2) and the metric tensor for polar coordinates will be
computed in two ways. It was shown in (3.4.1) and (3.4.2) that

-rsinθ cosθ
S = ⎛⎝ rcosθ sinθ ⎞⎠ = [ e1, e2 ] e1 = r(-sinθ, cosθ) e2 = (cosθ, sinθ) . (5.13.10)

One way to compute ḡ is this, using (5.13.9): ( 1=θ, 2=r )

T ⎛ -rsinθ rcosθ ⎞ ⎛-rsinθ cosθ ⎞ ⎛ r2 0 ⎞

ḡ = S S = ⎝ cosθ sinθ ⎠ ⎝ rcosθ sinθ ⎠ = ⇒ ḡ θθ = r2 ḡ rr = 1
⎝ 0 1⎠
hθ = r hr = 1
Another method uses (5.11.3) ( but in Picture C, so no prime on ḡ ),
(5.13.11)
⎛ e1•e1 e1•e2 ⎞ ⎛ r 0 ⎞
2
ḡ = ⎜ ⎟ = // det(ḡ) = r2 .
⎝ e2•e1 e2•e2 ⎠ ⎝ 0 1 ⎠

Notice that this metric tensor is in fact symmetric, and that one of its elements is a function of the
coordinates. The length2 of a small vector dx can be written using (5.2.5) (but in Picture C),

(ds)2 = ḡ km dxk dxm = ḡ θθ dθ dθ + ḡ rr dr dr = r2 (dθ)2 + (dr)2 . (5.13.12)

The Jacobian is given by (5.12.6) ( but in Picture C so J(x') → J(x)),

-rsinθ rcosθ
J(r,θ) = det(S(r,θ)) = det⎛⎝ cosθ sinθ ⎞⎠ = -r so |J| = r and g = J2 = r2, g =r. (5.13.13)

Note that J2 = g'/g → g/1 = g converting (5.12.14) to Picture C.

60
 Chapter 5: The Metric Tensor

Example 2: Spherical coordinates: metric tensor and Jacobian

As with Example 1, Picture C is used, this time with (x1, x2, x3) = (r,θ,φ) .

In (3.4.4) it was found that

⎛ sinθcosφ rcosθcosφ -rsinθsinφ ⎞

S = ⎜ sinθsinφ rcosθsinφ rsinθcosφ ⎟ . (3.4.4)
⎝ cosθ -rsinθ 0 ⎠

Comment: The matrix S looks the same in all Pictures. What differs in different Pictures are the names of
the coordinates that go with the rows and columns of the matrix. Looking at Picture A of Fig (2.1.1) and
equations (2.1.6), one sees that the matrix elements of S in Picture A are Sik = (∂xi/∂x'k) where xi are the
Cartesian coordinates x,y,z and x'i are the curvilinear coordinates r,θ,φ. But in Picture C of Fig (5.13.8),
one has instead Sik = ∂ξi/∂xk where ξi are the Cartesian coordinates x,y,z and xi are the curvilinear
coordinates r,θ,φ.

The metric tensor from (5.13.9) is then given by a Maple matrix calculation as

⎛ 1 02 0 ⎞
T
ḡ = S S = ⎜ 0 r 0 ⎟ det(ḡ) = r4sin2θ (5.13.14)
⎜ 0 0 r2sin2θ ⎟
⎝ ⎠

so that

ḡ 11= ḡ rr = 1 h1 = hr = ḡ rr = 1
2
ḡ 22= ḡ θθ = r h 2 = hθ = ḡ θθ = r
2 2
ḡ 33= ḡ φφ = r sin θ h 3 = hφ = ḡ φφ = rsinθ . (5.13.15)

The Jacobian using (5.12.6) is found by Maple to be,

J(r,θ,φ) = det(S) = r2sinθ . (5.13.16)

An alternate calculation uses (5.12.14) converted to Picture C where J2 = g'/g → g/1 = g. Then,

J = g = det(ḡ) = r4sin2θ = r2sinθ. (5.13.17)

Differential distance using (5.2.5) (but in Picture C) and (5.13.15) is:

(ds)2 = ḡ km dxk dxm = (dr)2 + r2(dθ)2 + r2sin2(dφ)2 (5.13.18)

and if dφ = 0, this agrees with the polar coordinates result (5.13.12).

61
 Chapter 5: The Metric Tensor

5.14 Special Relativity and its Metric Tensor: vectors and spinors

In this Section the Standard Notation introduced below in Chapter 7 is used. In that notation Rij is
written Rij , contravariant vectors Vi are written Vi, and covariant vectors V̄j are written Vj. It is a
tradition in special and general relativity to use Greek letters for 4-vector indices and Latin letters for
spatial 3-vector indices.
The (Quasi-Cartesian) metric tensor of special relativity is frequently taken as G = diag(1,-1,-1,-1)
and the ordering of 4-vectors as xμ = (t,x,y,z) where c=1 (speed of light) and μ= 0,1,2,3 ( Bjorken and
Drell p 281). General relativity people often use G = diag(-1,1,1,1) ≡ η instead (Weinberg p 26). Still
other authors use G = 1 and xμ = (it,x,y,z) where i is the imaginary i, but this approach does not easily fit
into our tensor framework which is based on real numbers.
A Lorentz transformation is a linear transformation Fμν (equation below has an implied sum on ν)

x'μ = Fμν xν = Rμν xν ⇒ xν is a contravariant vector // x' = F(x) (5.14.1)

and a theory requirement is that invariant length be preserved, x'.x' = x.x = scalar. Special relativity also
requires that the metric tensor G be the same in all frames, since no frame is special, so G' = G. But this
says, in our old notation, that R G RT = G (which is (5.7.6) with g' = g = G). This condition restricts the
(proper) Lorentz transformations to be rotations, boosts (velocity transformations), or any combination of
the two. In particular,

R G RT = G ⇒ det(R G RT) = det(G)

⇒ det(R) det(G) det(RT) = det(G) ⇒ [det(R)]2 (-1) = (-1)
⇒ det(R) = ±1 . (5.14.2)

"Proper" Lorentz transformations have det(R) = det(F) = +1, and here are two examples. First, a boost
transformation in the x direction,

⎛ cosh(b) sinh(b) 0 0⎞ ⎛ 0 i 0 0⎞
⎜ ⎟ ⎜ ⎟
μ ⎜ sinh(b) cosh(b) 0 0⎟ ⎜ i 0 0 0⎟
F v= ⎜ = exp(-ibK1) where (K1)μν =⎜ (5.14.3)
0 0 1 0⎟ 0 0 0 0⎟
⎜ ⎟ ⎜ ⎟
⎜ 0 0 0 1 ⎟⎠ ⎜ 0 0 0 0⎟
⎝ ⎝ ⎠

and second, a rotation transformation about the x axis,

⎛1 0 0 0 ⎞ ⎛0 0 0 0⎞
⎜ ⎟ ⎜ ⎟
μ ⎜0 1 0 0 ⎟ μ ⎜0 0 0 0⎟
F v= ⎜ = exp(-irJ1) where (J1) ν = ⎜ . (5.14.4)
0 0 cos(r ) − sin( r ) ⎟ 0 0 0 − i⎟
⎜ ⎟ ⎜ ⎟
⎜0 0 sin( r ) cos(r ) ⎟⎠ ⎜0 0 i 0 ⎟⎠
⎝ ⎝

The matrices K1 and J1 are called generators and are a part of a set of six 4x4 matrices Ji and Ki for i =
1,2,3. These 6 generator matrices satisfy a set of commutation relations known as a Lie Algebra,

62
 Chapter 5: The Metric Tensor

[ Ji, Jj] = +i εijkJk // [ A,B ] ≡ AB - BA

[ Ji, Kj] = +i εijk Kk
[ Ki, Kj] = -i εijkJk . (5.14.5)

In these commutators, the generators Ji and Ki can be regarded as abstract non-commuting operators,
while the specific 4x4 matrices shown above for J1 and K1 are just a "representation" of these abstract
operators as 4x4 matrices. The six 4x4 generator matrices Ji and Ki are (g = G = diag(1,-1,-1,-1) )

(Jμν)αβ = i ( gμαδνβ – gναδμβ) (J1)αβ ≡ (J23)αβ = i ( g2αδ3β – g3αδ2β) and cyclic 123
(K1)αβ ≡ (J01)αβ = i ( g0αδ1β – g1αδ0β) and cyclic 123 (5.14.6)

where (-i)(Jμν)αβ = ( gμαgνβ – gναgμβ) is a rank-4 tensor, antisymmetric under μ↔ν and α ↔ β.

An arbitrary Lorentz transformation can be represented as Fμv(r,b) = [ exp {– i ( r•J + b•K)} ]μν where
the 6 numbers r and b are called parameters (rotation and boost) and this F is a combined boost/rotation
transformation (note that eA+B ≠ eAeB for non-commuting matrices A,B). The product of two such Lorentz
transformations is also a Lorentz transformation, and in fact the transformations form a continuous group
known as the Lorentz Group, which then has 6 parameters.
The first two commutators shown above (all J and all K) are each associated with a 3 parameter
continuous group called the rotation group. The abstract generators of this group can be "represented" as
matrices of any dimension, and are labeled by a number j such that 2j+1 is the matrix dimension. For
example, the 2x2 matrix representation of the rotation group is labeled by j = 1/2, and is called the spinor
representation and is associated in physics with the "intrinsic spin" of particles of spin 1/2 such as
electrons. The vectors (spinors) in this case have two elements, and (1,0) and (0,1) are "up" and "down".
Representations of the Lorentz group have labels {j1, j2}, where j1 is for the J-generated rotation
subgroup, and j2 for the K-generated rotation subgroup, and are usually denoted j1⊗j2. Such a
representation then has vectors containing (2j1+1)(2j2+1) elements. In the case 1/2⊗1/2 there are 2*2=4
elements in a vector, and when these elements are linearly combined in a certain manner, they form the 4-
vector object which one writes as Aμ such as xμ. This is the "vector representation" of the Lorentz group
upon which is built the entire edifice of special relativity tensor algebra.
Two other basic representations of the Lorentz group are these: 1/2 ⊗ 0 and 1/2 ⊗ 0. These are 2x2
matrix representations and they are different 2x2 representations. For each representation one can
construct a whole tensor analysis world based on 2-vectors. Just as with the 4-vectors, one has
contravariant and covariant 2-vectors. The two representations 1/2 ⊗ 0 and 1/2 ⊗ 0 are called spinor
representations since they are each 2-dimensional. Since there are two distinct spinor representations, one
needs some way of distinguishing them from each other. One representation might be called "undotted"
and the other "dotted" and then there are four 2-vector types to worry about, which transform this way,

• • •
V'a = RabVb V'a = Rab•Vb contravariant 2-vectors
b •
V'a = Ra Vb V'•a = R•abV•b covariant 2-vectors (5.14.7)

where now dots on the indices indicate which Lorentz group representation that index belongs to. The 2x2
•
matrices Rab and Ra•b are not the same. A typical rank-2 tensor would transform this way,

63
 Chapter 5: The Metric Tensor

• • •
X' ab = Raa'Rb•b' Xa'b' . (5.14.8)

This then is the subject of what is sometimes called Spinor Algebra as opposed to Tensor Algebra, but it
is really just regular tensor algebra with respect to the two spinor representations of the Lorentz group.
We have inserted this blatant digression just to show that the general subject of tensor analysis
includes all this spinor stuff under its general umbrella.
In closing, Maple shows that the metric tensor G is indeed preserved under boosts and rotations. In
Maple,

evalm(Bx &* G &* transpose(Bx)) means Bx G BxT

and Maple is just verifying that Bx G BxT = G and similarly Rx G RxT = G :

64
 Chapter 5: The Metric Tensor

(5.14.9)

5.15 General Relativity and its Metric Tensor

In general relativity a Picture of interest is Picture C

x(0) → ξ, g(0) → η (5.15.1)

but the x(0)-space is replaced by a Quasi-Cartesian [see (1.10)] space with coordinates ξi with the metric
tensor of special relativity which is now called η.
This ξ-space represents a "freely-falling" coordinate system in which the laws of special relativity
apply and the metric tensor is taken to be G = Ḡ = diag(-1,1,1,1) ≡ η.
The xi are the coordinates of some other coordinate system. There is some transformation x = F(ξ)
which defines the relationship between these two systems. The covariant metric tensor in x-space is
written ḡ = STGS = STηS. This is (5.7.6) for Picture C with g(0) = G = η, "quasi-Cartesian". Using the
Standard Notation introduced in Chapter 7 below, this is usually written as

ḡ = STηS // translation of (5.7.6) to our quasi-Cartesian Picture C

gdn = ST ηdnS // developmental notation, gdn has "down" indices

gμν = (ST)μα ηαβ Sβν = Sαμ ηαβ Sβν // Standard Notation as in Chapter 7

65
 Chapter 5: The Metric Tensor

gμν = (∂ξα/∂xμ) ηαβ (∂ξβ/∂xν) // e.g., Sαμ = (∂ξα/∂xμ)

gμν(x) = (∂ξα/∂xμ) (∂ξβ/∂xν) ηαβ // Weinberg p 71 (3.2.7) . (5.15.2)

The last line then defines the gravitational metric tensor in x-space based on the transformation
ξ = F-1(x) = ξ(x). ( This and the following references are from the book of Weinberg, see References.)
Newton's Second Law ma = f appears this way in general relativity,

m (∂2xμ/∂τ2) = fμ - m Γμνλ (∂xν/∂τ) (∂xλ/∂τ) // p 123 (5.1.11 following) (5.15.3)

where fμ is an externally applied force, but there is then an extra bilinear velocity-dependent term which
represents an effective gravitational force (it acts on mass m) arising from spacetime itself. The object
Γμνλ is called the affine connection and is related to the metric tensor in this way.

Γμνλ = (1/2) gμσ( ∂νgλσ + ∂λgνσ – ∂σgνλ ) . // p 75 (3.3.7) (5.15.4)

These very brief comments are only meant to convince the reader that the equations of general relativity
also have their place under the general umbrella of tensor analysis as discussed in this document. The fact
that Γμνλ is not a mixed rank-3 tensor is demonstrated for example in (F.6.3).

5.16 Continuum Mechanics and its Metric Tensors

One can describe (Lai) the forward "flow" of a continuous blob of matter by x = x(X,t) where X =
x(X,t0). A "particle" of matter (imagine a tiny cube) that starts at location X at time t0 ends up at x at time
t. Two points in the flow separated by dX at t0 end up separated by some dx at t. The relation between
these separation vectors is given by dx = F dX where F is called the deformation gradient. F describes
how a particle starting say with a cubic shape at t0 gets deformed into some parallelepiped (3-piped)
shape at t as shown in Fig (5.16.7) below. If we examine dx = F dX we find that | dx | ≠ | dX | since the
vector dX typically gets rotated and stretched as dX → dx during the flow. [ This flow is further described
in Appendix K. ]
The finite-duration flow x = x(X,t) from time t0 to time t can be thought of as a (generally non-linear)
transformation of the form x = F(X,t), where we can regard t as an added parameter. Recall from Chapter
2 that a general transformation was annotated as x' = F(x) or x = F-1(x') and the linearized transformation
was dx' = R dx or dx = S dx', as in (2.1.6). To be compatible with Lai notation which uses symbol F for
the deformation gradient, and to make things clearer, we shall replace our Chapter 2 transformation name
F by the name F-1, and we shall identify x'-space with X-space. We then have this modified version of
Picture A,

66
 Chapter 5: The Metric Tensor

(5.16.1)

In this flow we assume Cartesian coordinates in both x-space and X-space, so the metric tensors which
determine physical distance in these spaces are both 1. Before continuing, it is helpful to have an example
of a fluid flow situation.

Example: A 2D Fluid Flow based on the Polar Coordinates Transformation

We first quote Fig (3.4.3) with a few enhancements ( F ≡ F-1) :

(3.4.3) (5.16.2)

The tiny white square on the left maps into a tiny white 2-piped shown on the right under this
transformation. The edges of this 2-piped in x-space (on the right) are dr and rdθ, there is no confusion
about that, and the area is dA = rdrdθ. The question at hand is how one should interpret the edges of the
tiny white square on the left, and what is the x'-space metric tensor ḡ' on the left? It is shown in Section
C.5 that there are two distinct interpretations and each has its own use. We call the two interpretations the
Cartesian View and the Curvilinear View.
⎛ r2 0 ⎞
The view we are familiar with is the Curvilinear View where ḡ' =
⎝ 0 1 ⎠ . In this view, the top
edge of the white square on the left has length ds' = rdθ and this matches the corresponding edge of the 2-
piped on the right, which is ds = rdθ. This is the view in which hypothesis (5.2.6) is valid (that ds' = ds),
so all our tensor machinery is applicable, such as the fact (5.7.9) that ḡ' = STS. One might call this the
covariant view, since distance ds is a scalar under the transformation. Here is a bit more detail.
For the upper edge of the white square on the left, we can set dx' = dθ e'θ= (dθ r) ^ e 'θ. This maps into
the northeast edge of the 2-piped on the right which is then dx = (rdθ) ^e θ. One then verifies that

67
 Chapter 5: The Metric Tensor

(ds')2 = dx' • dx' = g'θθ dθ dθ = r2 dθ dθ = (rdθ)2 // Curvilinear View

(ds)2 = dx • dx = (rdθ) ^
e θ • (rdθ) ^
e θ = (rdθ)2 (5.16.3)

10
In the Cartesian View, we instead use ḡ' = ⎛⎝ 0 1 ⎞⎠ on the left. In this case, the edges of the white
square are dr and dθ, just as if these were regular Cartesian coordinates. As explained in Appendix C, one
advantage of the Cartesian View is that, for a general coordinate system, it is the only view that can be
drawn on a piece of paper (2D) allowing the reader to have some comprehension of what is going on (and
this is even more true in 3D). A disadvantage of the Cartesian View is that ds ≠ ds' so it is "non-
covariant". In the Cartesian View we find, for our same top edge vector,

(ds')2 = dx' • dx' = 1 dθ dθ = (dθ)2 // Cartesian View

(ds)2 = dx • dx = (rdθ) ^
e θ • (rdθ) ^
e θ = (rdθ)2 (5.16.4)

and so ds ≠ ds' which says | dx | ≠ |dx'|.

To arrive at our Example model for a fluid flow, we redraw Fig (5.16.2) renaming x' = (θ,r) to be X =
(X,Y), and make corresponding changes elsewhere in the figure:

(5.16.5)

The tiny white square on the left is a blob of (2D) fluid at time t = 0 and after flowing a while it ends up
in the location of the white 2-piped on the right at time t = 1. This is just a simple 2D flow example
making use of a transformation we are already familiar with (see Comment below). For a general 3D
forward flow transformation x = F(X,t) one finds that the differential white cubic blob at t = 0 ends up in
a 3-piped at t = t which is both rotated and stretched relative to the starting cube. Due to this stretch, one
finds in general that |dx| ≠ |dX| where these are vector lengths relative to the Cartesian View metric
tensor, which is in fact the metric tensor that determines physical distance in both X-space and x-space.
10
Thus, the ḡ' = ⎛⎝ 0 1 ⎞⎠ metric tensor for X space has this well-defined meaning -- it relates to actual

⎛ r2 0 ⎞ T
physical distance in X-space. However, it is the curvilinear view metric tensor ḡ' =
⎝ 0 1 ⎠ = S S whose

68
 Chapter 5: The Metric Tensor

matrix elements determine such things as how much edges stretch in the flow and how much areas and
volumes change during the flow, which we shall examine below.

Comment: The Example presented above really makes no flow sense since there is no time parameter in
the flow transformation! The purpose of the Example is merely to illustrate the idea of two metric tensors
in X-space, and to do so with a transformation that is familiar from our "curvilinear coordinates" study,
and to tie in the notions of the two Views from Appendix C. The Example could be made more
reasonable using a contrived transformation like this,

general t t=0 t=1

x = F(X,t): x = (1-t)X + t Y cos(X) x=X x = Ycos(X)
y = (1-t)Y + t Y sin(X) y=Y y = Ysin(X) . (5.16.6)

In this case, at t = 0 we have the correct x = F(X,t=0) = X and at the end of the flow (t = 1) we have the
transformation equations shown in Fig (5.16.5).

Having worked through this contrived example, we can now draw a figure showing the general mapping
of a differential blob during a flow from time t = 0 to time t = t. This figure is taken from Chapter 8 where
the X-space left side is called x'-space as in (15.6.1) and the initial "blob" is shown drawn in the Cartesian
View: (this picture happens to use the Chapter 7 Standard Notation for coordinates, contravariant = upper
index)

Flow X-space (= x'-space) t = 0 Flow x-space t = t

(8.2.2) (5.16.7)

We might be interested in knowing how much the dx'3 edge is stretched going from time t=0 to time t=t,
and by how much the volume changes, and so on. All these geometric questions are posed and answered
in Chapter 8, and we shall use some of those results below.

69
 Chapter 5: The Metric Tensor

Here then is a translation table comparing the continuum mechanics notation of Lai to the tensor notation
used earlier in our document (and in the above Example).

continuum mechanics our document ( Forward Flow X → x )

X, x ↔ x', x
x = x(X,t) = F(X,t) ↔ x = F(x',t) = F-1(x',t) // Lai p 70 (3.1.4)
dx = F dX ↔ dx = S dx' // as in (2.1.6) // Lai p 86 (3.7.6) or p 105 (3.18.3)
F ↔ S
F-1 ↔ R // R = S-1
x = Cartesian ↔ ḡ = 1
X = Cartesian ↔ ḡ ' = 1 // Cartesian View
C = FTF ↔ ḡ' = STS // Curvilinear View, as in (5.7.9) // Lai p 114 (3.23.2)
C-1 = F-1(F-1)T ↔ T -1
g' = RR // since g' = ḡ' and S = R -1
, as in (5.7.9)
(5.16.8)

Thus, the deformation gradient F is just the S matrix associated with transformation F = F-1. The
Curvilinear View metric tensor ḡ' = STS appears in Lai as C = FTF which is known as the right Cauchy-
Green deformation tensor (manifestly symmetric, so a viable metric tensor). [The left Cauchy-Green
deformation tensor is B = FFT ].

Given the above flow situation, it is then possible to add two more transformations F1 and F2 which take
X-space and x-space to independent sets of curvilinear coordinates X' and x': ( a new x' here)

(5.16.9)

and we then have an interesting triple application of the notions of Chapter 2 to a real-world situation.
This drawing is the implicit subject of Section 3.29 (p131-138) of Lai. The flow transformation of interest
here is

x' = F2(x) = F2(F(X)) = F2(F(F1-1(X'))) ≡ G(X') . (5.16.10)

In (reverse) dyadic notation the deformation gradient is written F = (∇x) where ∇ means ∇(X)so that

dx = F dX = (∇x) dX Fij = (∇x)ij = ∂j(X)xi = ∂xi/∂Xj . (5.16.11)

70
 Chapter 5: The Metric Tensor

The (∇x) notation is explained in (E.4.4), and in Appendix G the object (∇v) for an arbitrary vector field
v(x) is expressed in general curvilinear coordinates.

Consider again Fig (5.16.7) displayed above:

Flow X-space (= x'-space) t = 0 Flow x-space t = t

(8.2.2) (5.16.7)

We can identify the mapping shown in this picture with our Flow situation of table (5.16.8). Chapter 8
discusses in much detail how length, volume and area transform under a general transformation. The
length, area and volume magnitudes on the left are called dL'n = dx'(n), dA'n and dV', while the
corresponding quantities on the right are called dx(n), d¯A(n) and dV, where the first two items are
vectors. Cribbing the results of (8.4.g.2) and converting them from standard notation to developmental
notation, we have

| dx(n)|/ dL'n = h'n = [ḡ'nn]1/2 = the scale factor for edge dx(n)

| d¯A(n)|/ dA'n = (1/h'n) |J| = (1/h'n) g'1/2 = [g'nn g']1/2 = [cof(ḡ'nn)]1/2

|dV| / dV' = |J| = g'1/2 // where g' ≡ det(ḡ'ij) = J2 , ḡ' = STS (8.4.g.2) (5.16.12)

We can then translate these three lines into our Flow context:

| dx(n)| / | dX(n)| = h'n = [ḡ'nn]1/2 = [(FTF)nn]1/2 = [Cnn]1/2 // Lai p 114 (3.23.6-8)

dAn| / |dA0n| = [g'nng']1/2 = [cof(ḡ'nn)]1/2 = [cof(FTF)nn)]1/2 = [cof Cnn)]1/2 // Lai p 129 (3.27.11)*

|dV| / |dV0| = |J| = g'1/2= [det(ḡ'ij)]1/2 = [det(FTF)]1/2 = |det(F)| // Lai p 130 (3.28.3) (5.16.13)

where

71
 Chapter 5: The Metric Tensor

edge area volume

(n)
X-space : dX d A0 n dV0 time t0
x-space : dx(n) dAn dV time t (5.16.14)

Thus, for example, the volume change of a "flowing" particle of continuous matter is given by the
Jacobian |J| = |detF| associated with the deformation gradient tensor F. We put quotes on "flowing" only
because this might be a particle of solid steel that is momentarily moving and deforming a very small
amount during an oscillation or in response to an applied stress.

In the middle line of (5.16.13) we state that | dAn| / | dA0n| = [cof(FTF)nn)]1/2 and quote Lai p 129
(3.27.11) for verification. However, what Lai (3.27.11) actually says (slightly translated to our notation) is
this:

dA(n)/dA(n)0 = det(F) | (F-1)T un | un = unit base vector, (un)i = δn,i (5.16.15)

which seems a far cry from our result [cof(FTF)nn)]1/2. But consider, using F-1 = R from table (5.16.8),

| (F-1)T un |2 = | RT un |2 = [RTun]i[RTun]i = Rni Rni = (RRT)nn = g'nn (5.16.16)

so
det(F) | (F-1)T un | = g'1/2 g'nn1/2 = [cof(ḡ'nn)]1/2 = [cof(FTF)nn)]1/2 . (5.16.17)

Here det(F) = det(S) = g'1/2 from (5.12.19) and we have used Theorem 1 (8.4.f.1) converted to
developmental notation,

g' g'nn = [cof(ḡ'nn)] . // more generally, (detA) A-1 = cof(A) if A = AT (5.16.18)

It might be noted that the Lai book does in fact use our "developmental notation" in that all indices are
written "down" (when indices are shown), but no overbars mark covariant objects. Here are a few
examples:

Lai notation Developmental notation Standard Notation

dA0 = dX(1)x dX(2) (3.27.1) d¯A0 = dX(1)x dX(2) (dA0)i= εijk [dX(1)]j [dX(1)]k
[divT]i = ∂jTij (4.7.3) [divT]i = ¯
∂ jTij [divT]i = ∂jTij
(5.16.19)

Of course when Cartesian coordinates are assumed, the up and down position makes no difference.

Lai writes tensors in bold face such as F for the deformation gradient noted above, or T for the stress
tensor. Perhaps this is done to emphasize the notion of a tensor as an operator, as in our Section E.7. Lai
writes a specific matrix as [T], but a matrix element is Tij. Notation is an ongoing burden and each area
of physics seems to have its own accepted conventions.

72
 Chapter 6: Reciprocal Base Vectors

6. Reciprocal Base Vectors En and Inverse Reciprocal Base Vectors U'n

6.1 Definition of the En

This entire Chapter uses the Picture A context,

(6.1.1)

Although various definitions are possible, we shall define the reciprocal tangent vectors En in the
following manner [ implied sum on i, and en = ∂'nx from (3.2.6)]

En ≡ g'ni ei = g'ni ∂'ix ⇒ en = ḡ'niEi . // since ḡ' = g'-1 (6.1.2)

Comment: Notice how this differs in structure from the rule for forming a covariant vector from a
contravariant one,

V̄n = ḡ'niVi . (6.1.3)

In the last equation, the right side is a linear combination of vector components Vi, while in the previous
equation (6.1.2) the right side is a linear combination of vectors ei. In this case, i is a label on ei ,
whereas in the other case i is an index on Vi. Labels and indices are different.

Since the tangent base vectors ei are contravariant vectors in x-space by (3.3.2), and since En is a linear
combination of the ei, the En are also contravariant vectors in x-space. Notice that in the definition
En ≡ g'ni ei, these two x-space vectors are related by the metric tensor of the other space, x'-space.

One can express the components of En in two ways,

(En)k ≡ g'ni (ei)k = Skig'ni // (ei)k ≡ Ski by (3.2.5)

= g'niSki = Rnigik // Rnigik = g'ni(ST)ik = g'niSki by (5.7.8) Rg = g' ST (6.1.4)

Taking i→c and then k→i (and g is symmetric),

(En)i = g'ncSic = gicRnc . // sum on second indices (6.1.5)

Applying R to both sides of En ≡ g'ni ei in (6.1.2) gives En transformed into x'-space,

E'n = g'nk e'k

so
(E'n)i = g'nk (e'k)i = g'nkδk,i = g'ni . (6.1.6)

73
 Chapter 6: Reciprocal Base Vectors

Using the axis-aligned basis vectors un in x-space, (un)m = δn,m, we also have this dot product,

En • um = Rnm . (6.1.7)

Proof: En • um = ḡ ab(En)a(um)b = ḡ ab (gacRnc)(δm,b) = (ḡ magac) Rnc = δm,c Rnc = Rnm .

6.2 The en and En Dot Products and Reciprocity (Duality)

Three covariant dot products are of great interest. The first is this (see (5.7.6) for last step)

en • em = ḡ ij (en)i (em)j = ḡ ij Sin Sjn = STni ḡ ij Sjm = ( ST ḡ S)nm = ḡ'nm . (6.2.1)

The second is

En • em = ḡ ij (En)i (em)j = ḡ ij gic Rnc Sjm = δj,c Rnc Sjm = Rnj Sjm = (RS)nm = δn,m (6.2.2)

and the third is

En • Em = ḡ ij (En)i (Em)j = ḡ ij gic Rnc gjb Rmb = δj,c Rnc gjb Rmb = Rnj gjb Rmb

= Rnj gjb RTbm = (R g RT)nm = g'nm . // last step is (5.7.6) (6.2.3)

To summarize the last three results and (6.1.2),

en • em = ḡ'nm ⇒ |en| = ḡ'nn = h'n (scale factor) En = g'ni ei

En • em = δn,m en = ḡ'ni Ei
En • Em = g'nm ⇒ |En| = g'nn . (6.2.4)

When the relation between two sets of vectors en and Em is as shown in (6.2.4), the vectors are said to be
reciprocal or dual to each other. Since the en already have the name "tangent base vectors", we refer to
the En as the "reciprocal base vectors". The notion of reciprocal or dual vector sets is studied in more
detail in the notes just below.
Notice that the two relations on the right in (6.2.4) are implied by the three dot product relations on
the left. For example, since the en are a complete set, we can expand En with temporarily unknown
coefficients αnk to get En = αnkek. Dotting both sides with Em then reveals the coefficients:

g'nm = En • Em = (αnkek) • Em = αnk (ek • Em) = αnk δk,m = αnm

and therefore we find En = αnkek = g'nkek = g'niei as shown top right in (6.2.4).

74
 Chapter 6: Reciprocal Base Vectors

Whereas en and Em exist in x-space, e'n = Ren and E'm = REm exist in x'-space -- these are just the
transformed contravariant vectors using the rule (2.3.2), v' = R(x)v. Recall from (3.2.1) that the e'n are
axis-aligned unit vectors in x'-space, (e'n)i = δn,i.

Using the fact (6.1.6) that (E'n)i = g'nk (e'k)i we find that,

E'n • e'm = ḡ'ij (E'n)i(e'm)j = ḡ'ij g'ni δm,j = ḡ'im g'ni = δn,m . (6.2.5)

This is consistent with (6.2.4) and (5.10.2) that A' • B' = A • B,

E'n • e'm = En • em = δn,m . (6.2.6)

The other two dot products above work this same way. We can then rewrite (6.2.4) for the primed vectors,

e'n • e'm = ḡ'nm ⇒ |e'n| = ḡ'nn = h'n (scale factor) E'n = g'ni e'i
E'n • e'm = δn,m e'n = ḡ'ni E'i
E'n • E'm = g'nm ⇒ |E'n| = g'nn . (6.2.7)

As an exercise, we can verify the top left equation in this manner, knowing (e'n)i = δn,i,

e'n • e'm = ḡ'ij (e'n)i (e'm)j = ḡ'ij δn,i δm,j = ḡ'nm

The vectors en and En are examples of the general notion of dual vector sets which we now describe.

Notes on Reciprocity (Duality)

1. Suppose some set of vectors bn forms a complete basis for x-space. Can one find a set of vectors Bn
that have the property

Bm • bn = δm,n ? // duality relation; Ba and ba are reciprocal (6.2.8)

As shown below, the answer is normally "yes", and the vectors Bn are uniquely determined by the bn.
One says that the set {Bn} is "dual to" the set {bn} and vice versa. If we regard bn as a basis, then Bn is
the "dual basis", and Bm • bn = δm,n is the "duality relation". Another terminology is that the vectors Bn
are "reciprocal to" the vectors bn and vice versa.

2. It was shown in (6.2.2) that En • em = δn,m so the En and the en vectors are dual to each other. The en
are the tangent base vectors, and the En are "reciprocal to" the en which is why we call them the
reciprocal base vectors.

3. One can solve for the Bn in terms of the bn. Each Bn has N components, so there are N2 unknowns.
The duality relation Bm • bn = δm,n is a set of N2 equations. This is basically a Cramer's Rule problem in

75
 Chapter 6: Reciprocal Base Vectors

N2 variables. Since the bn form a complete basis, one can expand Bm on the bn with some coefficients we
will call w'mn ( at this point w'mn is unknown),

Bm = w'mnbn . (6.2.9)

Then from (6.2.8),

δm,k = Bm • bk = w'mn bn • bk . // bn • bk = ḡ ij(bn)i(bk)j . (6.2.10)

Define matrix W' by,

W'nk ≡ bn • bk (6.2.11)

and note that W'nk is symmetric. Then (6.2.10) says

δm,k = w'mnW'nk or w'W' = 1 or w' = W'-1 . (6.2.12)

Assuming for the moment that detW' ≠ 0, the solution is given by w' = W'-1 . The (5.6.4) "Digression"
showed that (A-1)T = (AT)-1 for invertible A, so w'T = (W'-1)T = (W'T)-1 = W'-1 = w' and therefore w' is
symmetric as well. Since W' is known from (6.2.11), w' = W'-1 and the Bm = w'mnbn of (6.2.9) have been
found. Finally,

Bm • Bn = w'mibi • w'njbj = w'miw'nj bi • bj = w'miw'njW'ji = w'miδni = w'nm . (6.2.13)

4. In the case that bn = en and Bn = En, one finds that W'nk = en • ek = ḡ'nk , the covariant metric tensor.
Then w' = W'-1 must be the contravariant metric tensor w'mn = g'mn = Em • En. Then Bm = w'mnbn says
En = g'mn en which agrees with our original definition (6.1.2) of the En. Moreover, det(W') = det( ḡ'ab) =
g' of (5.12.14), so as long as g' ≠0, one has detW' ≠ 0.

5. For the general bn case, one can imagine that the bn are the tangent base vectors for some
transformation x" = Fb(x) (with some linearized Rb ≠ R where R goes with F), then the issue of detW' ≠ 0
boils down to g' ≠ 0 for that transformation. In the case that x-space is Cartesian, we know that J2 =
det(ḡ')/ det(ḡ) = g'/g = g'. If this transformation Fb is invertible, then J ≠ 0 everywhere, so g' ≠ 0 and
then detW ≠ 0.

6. If the bm are true tensorial vectors, then the Bm = w'mnbn will be as well (linear combination) and then
Bn • bm is a tensorial scalar. Therefore if Bn • bm = δn,m in x-space, then so also B'n • b'm = δn,m in x'-
space, where bm' = Rbm and B'n = RBn. For example, E'n • e'm = δn,m in x'-space where em' = Rem and
E'n = REn .

7. In Section 6.5 we shall encounter another dual pair Un • um = U'n • u'm = δn,m which is associated with
the inverse transformation x = F-1(x').

76
 Chapter 6: Reciprocal Base Vectors

8. One major significance of the equation Bn • bm = δn,m is that it allows the following expansions:

V = Σn kn Bn where km = V • bm
V = Σn cn bn where cm = V • Bm (6.2.14)

so that for example bm • V = bm • [Σn kn Bn] = Σn kn bm • Bn = Σn kn δm,n = km. These expansions are
explored in Section 6.6 below for the two dual sets En, en and Un, un.

9. Saying that the basis bn is complete implies that any vector V can be expanded in this basis. Consider
the second expansion shown in (6.2.14). Installing cn = V • Bn into the expansion one finds,

V = Σn (V • Bn) bn
so
Vb = Σn [ΣaVa(B̄n)a] (bn)b // from (5.10.3)

= ΣaVa { Σn( B̄n)a (bn)b } .

The last equation can only be valid if

Σn( B̄n)a (bn)b = δa,b // completeness relation for the dual set {bn, Bn } (6.2.15)

and this is the formal statement that the basis bn is complete.

We can verify this completeness claim for {en, En} as follows

Σn (Ēn)a(en)b = Σn (ḡ En)a(en)b = Σnc ḡ ac(En)c(en)b // from (5.8.1)

= Σnc ḡ ac [ Σk gckRnk] [Sbn] // from (6.1.5) and (3.2.6)

= Σnck SbnRnk gkc ḡ ca = [SRgḡ]ba = [ 1 1 ]ba = 1ba = δb,a . (6.2.16)

10. As a final step, we want to write the completeness relation (6.2.15) in matrix form. To do this, we first
take a short digression.

If b is a column vector, and if bT is the corresponding row vector version of b, then (bT)i = bi.

If a and b are vectors of dimension N, and if it happens that

(a)i(b)j = cij , (6.2.17)

then one can write

(a)i(bT)j = cij . (6.2.18)

77
 Chapter 6: Reciprocal Base Vectors

This equation can be written in matrix form as,

abT = c . (6.2.19)

Graphically, here is an example for dimension N=2 which shows how abT= c agrees with (a)i(b)j = cij :

abT =
⎛ a1 ⎞ (b b ) = ⎛ a1b1 a1b2 ⎞ = a matrix = ⎛ c11 c12 ⎞ . (6.2.20)
⎝ a2 ⎠ 1 2 ⎝ a2b1 a2b2 ⎠ ⎝ c21 c22 ⎠

This notion can be trivially generalized to apply to a sum of vectors. Let a[n} and b[n] each be a set of N
vectors labeled by n. Then the following component equation

Σn (a[n])i(b[n])j = cij (6.2.21)

can be written in matrix form as

Σn a[n] ( b[n])T = c . (6.2.22)

Now, since the completeness relation (6.2.15) has the form of (6.2.21)

Σn( B̄n)a (bn)b = δa,b // completeness relation for the dual set {bn, Bn } (6.2.15)

it can be expressed in the following compact matrix form from (6.2.22), where c = 1 is the identity matrix,

Σn B̄n bnT = 1 // completeness (matrix form) . (6.2.23)

In the special case that bn = en this reads

Σn Ēn enT = 1 // completeness (matrix form) . (6.2.24)

6.3 Covariant partners for en and En

A covariant partner vector is formed according to (5.8.4) which says V̄ = ḡ V or V̄i = ḡ ijVj, so

ē n = ḡ en ⇒ ( ē n)i = ḡ ij(en)j = ḡ ijSjn // (3.2.6)

= [ ḡS ]in = [RT ḡ']in = Rji ḡ'jn // (5.2.8) (6.3.1)

Ēn = ḡ En ⇒ (Ēn)i = ḡ ij(En)j = ḡ ijgjcRnc = δi,cRnc = Rni // (6.1.5) (6.3.2)

The components of the four vectors are then

(en)i = Sin (3.2.6) (En)i = gicRnc = g'ncSic (6.1.5)

( ē n)i = ḡ ijSjn = Rji ḡ'jn (6.3.1) (Ēn)i = Rni (6.3.2) (6.3.3)

78
 Chapter 6: Reciprocal Base Vectors

In (3.2.7) we wrote S = [e1, e2, e3 .... eN ] as a representation of the (en)j = Sjn , showing that the tangent
base vectors are the columns of matrix S. In the equation (Ēn)i = Rni the index order is not reversed as it
is in (en)j = Sjn, so one concludes that the Ēn as are the rows of matrix R

⎡ Ē ⎤
Ē1
2
R = ⎢ Ē ⎥ 3 = [Ē1, Ē2, Ē3 .... ĒN ]T (6.3.4)

⎣ Ē ⎦
...
N

Now recall how vectors transform,

V' = R V contravariant Rik(x) ≡ (∂x'i/∂xk) R = S-1

V̄' = ST V̄ covariant Sik(x') ≡ (∂xi/∂x'k) = STki(x') (2.5.1)

Here we regard both V and V̄ as existing in x-space, while V' and V̄' are the corresponding vectors in x'-
space.

As shown in the example (3.4.3), tangent base vectors en are vectors that one draws in x-space. They exist
in x-space. The relation En ≡ g'ni ei. of (6.1.2) says the En vectors are just linear combinations of the en
vectors, so the En vectors also exist in x-space and that is where one would draw them.

Based on this fact and the previous paragraph, we conclude the perhaps obvious fact that the vectors
en, ē n, En,Ēn all exist in x-space.

What are the corresponding vectors in x'-space that one gets from transformation x' = F(x) ? We just
apply the transformation rules quoted above to find out:

en' = Ren ⇒ (en')i = Rij (en)j = RijSjn = [RS]in = [1]in = δi,n (6.3.5)

in agreement with (3.2.1). Next,

En' = REn ⇒ (En')i = Rij (En)j = RijgjcRnc = [RgRT]in = g'in // (5.7.6) (6.3.6)

For the covariant vectors we find

ē n' = ST ē n ⇒ ( ē n')i = Sji ( ē n)j = Sji ḡ jkSkn = [ST ḡS]in = ḡ'in // (5.7.6) (6.3.7)

Ēn' = STĒn ⇒ (Ēn')i = Sji (Ēn)j = Sji Rnj = [RS]ni = δn,i (6.3.8)

The conclusions are

(en')i = δi,n (En')i = g'in

( ē n')i = ḡ'in (Ēn')i = δn,i (6.3.9)

79
 Chapter 6: Reciprocal Base Vectors

6.4 Summary of the basic facts about en and En

For use below, we package various results above into a single block. Here we use i,j for indices, and n.m
for basis vector labels,

( ē n)i = ḡ ij (en)j (Ēn)i = ḡ ij (En)j [ ē n = ḡ en Ēn = ḡ En] (5.8.4)

(e'n)i = Rij(en)j (E'n)i = Rij(En)j [ e'n = R en E'n = R En ]

( ē 'n)i = Sji( ē n)j (Ē'n)i = Sji(Ēn)j [ ē 'n = ST ē n Ē'n = ST Ēn ] (2.5.1)

(en)i = Sin (En)i = gijRnj = g'njSij (en')i = δi,n (En')i = g'ni

( ē n)i = ḡ ijSjn = Rji ḡ'jn (Ēn)i = Rni ( ē n')i = ḡ'ni (Ēn')i = δn,i
(6.3.3) (6.3.9)
en • em = ḡ'nm ⇒ |en| = ḡ'nn = h'n (scale factor) En = g'ni ei
En • em = δn,m en = ḡ'ni Ei
En • Em = g'nm ⇒ |En| = g'nn . (6.2.4)

e'n • e'm = ḡ'nm ⇒ |e'n| = ḡ'nn = h'n (scale factor) E'n = g'ni e'i
E'n • e'm = δn,m e'n = ḡ'ni E'i
E'n • E'm = g'nm ⇒ |E'n| = g'nn . (6.2.7)

(Ēn)i(en)j = δi,j [ Σn Ēn enT = 1 ] (6.2.16) and (6.2.24) (6.4.1)

6.5 Repeat the above for the inverse transformation: definition of the U'n

Section 3.5 introduced the inverse tangent base vectors called u'n . That is to say, the u'n are the tangent
base vectors of the inverse transformation x = F-1(x'). Whereas the forward tangent base vectors en exist
in x-space, the inverse tangent base vectors u'n exist in x'-space, so that is why we have put a prime on the
u'n .

Recall from (6.1.2) that En ≡ g'ni ei where En were the reciprocals of en . For the inverse transformation
we define in exact analogy (g ↔ g') some vectors U'n reciprocal to the u'n ,

U'n ≡ gni u'i ⇒ u'n = ḡ ni U'i . // since ḡ = g-1 (6.5.1)

This is just another application of the generic duality discussion of Section 6.2 with generic basis vectors
called bn and Bn. Everything goes along as in the previous Sections, but with the following changes to
account for the fact that we are now doing the inverse transformation x = F-1(x') :

g'↔ g R↔S en → u'n e'n → un En → U'n E'n → Un (6.5.2)

80
 Chapter 6: Reciprocal Base Vectors

and similarly for barred vectors. We can then manually translate the block of data presented in (6.4.1)
with the above rules to get:

(ū'n)i = ḡ'ij (u'n)j (Ū'n)i = ḡ'ij (U'n)j [ ū'n = ḡ' u'n Ū'n = ḡ' U'n]

(un)i = Sij(u'n)j (Un)i = Sij(Un')j [ un = S u'n Un = S Un' ]

(ūn)i = Rji(ū'n)j (Ūn)i = Rji(Ū'n)j [ ūn = RT ū'n Ūn = RT Ū'n ]

(u'n)j = Rjn (U'n)i = g'ijSnj = gnjRij (un)i = δi,n (Un)i = gni

(ū'n)i = ḡ'ijRjn = Sji ḡ jn (Ū'n)i = Sni (ūn)i = gni (Ūn)i = δn,i

u'n • u'm = ḡ nm ⇒ |u'n| = ḡ nn = hn (scale factor) U'n = gni u'i

U'n • u'm = δn,m u'n = ḡ ni U'i
U'n • U'm = gnm ⇒ |U'n| = gnn

un • um = ḡ nm ⇒ |un| = ḡ nn = hn (scale factor) Un = gni ui

Un • un = δn,m un = ḡ ni Ui
Un • Um = gnm ⇒ |Un| = gnn

(Ū'n)a(u'n)b = δa,b or Σn Ū'n u'nT = 1 (6.5.3)

It is helpful to keep all these eight vector symbol names in mind (and each has a covariant partner)

x'-space x-space (6.5.4)

axis-aligned basis vectors e'n un (e'n)i = δn,i (un)i = δn,i
dual partners to the above E'n Un (E'n)i = g'ni (Un)i = gni
tangent base vectors u'n en (u'n)i = Rin (en)i = Sin
reciprocal base vectors U'n En (U'n)i = g'iaSna (En)i = giaRna
= gnaRia = g'naSia

and recall that Xn • xm = δn,m for each of the four dual pairs (two primed, two unprimed).

6.6 Expanding vectors on different sets of basis vectors

x-space expansions on un and Un

Assume that V is some generic N-tuple V = (V1,V2....VN). There are various ways to expand V onto basis
vectors. One way is to expand on the axis-aligned basis vectors un, which recall live in x-space,

V = V1 u1 + V2 u2 +... = ΣnVnun where Un • V = Vn . (6.6.1)

The components Vn are Un • V because Un • um = δn,m. From (5.8.4) one can write Vn = gnmV̄m

81
 Chapter 6: Reciprocal Base Vectors

( regarded here as a definition of the V̄m) so one finds that

V = ΣnVnun = Σn gnmV̄m un = ΣnV̄m gmn un = ΣnV̄m Um (6.6.2)

and thus another expansion for V is this

V = V̄1 U1 + V̄2 U2 +.. = ΣnV̄nUn where un • V = V̄n . (6.6.3)

Comments:

1. If V is not a contravariant vector, one can still define V̄n = ḡ nmVm, but V̄n won't be a covariant vector.
A familiar example is that xn is never a contravariant vector if F is non-linear, but we can still talk about
the components x̄ n ≡ ḡ nmxm . In Standard Notation, xn → xn and x̄ n → xn and we do not hesitate to use
these two objects even though they are not tensorial vectors.

2. If V is a contravariant vector, the expansion above V = ΣnVnun displays the contravariant components
of V. The second expansion V = ΣnV̄m Um is still an expansion for contravariant vector V, but it displays
the components of the covariant vector V̄ which is the "partner" to V by V̄n = ḡ nmVm. It would be
incorrect to write this second expansion as V̄ = ΣnV̄m Um since that would say ḡ V = ΣnV̄m Um which is
just not true. We comment later on how this situation changes a bit in the Standard Notation.

x-space expansions on en and En

Another possibility is to expand V on the tangent basis vectors en, and we denote the components just
momentarily as αn,

V = α1 e1 + α2 e2 +... = Σn αn en . (6.6.4)

Using en • Em = δn,m one finds that

αn = En • V = (En)k Vk = Rnk Vk = V'n // ḡ = 1 so A•B = ḡ abAaBb = AkBk (6.6.5)

Therefore, the expansion is

V = V'1e1 + V'2e2 +... = Σn V'n en where En • V = V'n . (6.6.6)

If it happens that the N-tuple V = (V1,V2....VN) transforms as a contravariant vector, then Vn are the
contravariant components of that vector, and V'n are the contravariant components of V' in x'-space. On
the other hand, if V is not a tensorial vector, so Vn are not components of a contravariant vector, we can
still define V'n ≡ Rnk Vk, but then the V'n are not the contravariant components of V'.

Writing V'n = g'nmV̄'m the above expansion can be expressed as

82
 Chapter 6: Reciprocal Base Vectors

V = Σn V'n en = Σn,m g'nmV̄'m en = Σm V̄'m Σn g'mn en = Σm V̄'m Em (6.6.7)

so
V = V̄'1E1 + V̄'2E2 +... = Σn V̄'n En where en • V = V̄'n (6.6.8)

Summary of x-space expansions:

V = V1 u1 + V2 u2 +... = ΣnVn un where Un • V = Vn Un = gni ui

V = V̄1 U1 + V̄2 U2 +... = ΣnV̄n Un where un • V = V̄n
V = V'1 e1 + V'2 e2 +... = Σn V'n en where En • V = V'n En = g'ni ei
V = V̄'1 E1 + V̄'2 E2 +... = Σn V̄'n En where en • V = V̄'n (6.6.9)

Expanding on unit vectors. The covariant lengths of the different basis vectors are given by

|en| = |e'n| = g'

¯ nn |un| = |u'n| = ḡ nn
|En| = |E'n| = g'nn |Un| = |U'n| = gnn . (6.6.10)

Using these lengths, one can define unit vector versions of all the basis vectors and then rewrite the above
expansions as expansions on the unit vectors with lower case coefficients. For example, using

^
e n ≡ en/ g'
¯ nn (6.6.11)

the third expansion above becomes (script font for unit-vector components)

V = V'1^
e 1 + V'2^
e 2 +... = Σn V'n ^
en where ¯ nn En • V = V'n =
g' g¯'nn V'n . (6.6.12)

An example of a case where this last expansion would be useful is in the use of spherical curvilinear
coordinates, where for example ^
e1 = ^
r.

The N-tuple (V'1, V'2 ... V'N), although related to contravariant vector V (V'n = Rnk Vk), is not itself a
contravariant vector since it does not obey the rule V'n = Rnk Vk . In fact

V'n = Rnk Vk ⇒ (1/ g'

¯ nn ) V'n = Rnk (1/ ḡ kk ) Vk ⇒ V'n = Rnk ( g'
¯ nn / ḡ kk )Vk
(6.6.13)
x'-space expansions

Having done x-space expansions, we turn now to x'-space expansions. These can be obtained from the x-
space expansions by this set of rules,

g'↔ g R↔S u'n → en un → e'n U'n → En Un → E'n V'n ↔ Vn V̄'n ↔ V̄n

(6.6.14)
and here then are the x'-space expansions:

83
 Chapter 6: Reciprocal Base Vectors

V' = V'1 e'1 + V'2 e'2 +... = ΣnV'n e'n where E'n • V' = V'n E'n = g'ni e'i
V' = V̄'1 E'1 + V̄'2 E'2 +... = ΣnV̄'n E'n where e'n • V' = V̄'n
V' = V1 u'1 + V2 u'2 +... = Σn Vn u'n where U'n • V' = Vn U'n = gni u'i
V' = V̄1 U'1 + V̄2 U'2 +... = Σn V̄n U'n where u'n • V' = V̄n (6.6.15)

6.7 Another way to write the En

The reciprocal base vectors were defined above as linear combinations of the tangent base vectors, all in
the general Picture A context,

Ek ≡ g'ki ei . (6.7.1)

It is rather remarkable that there is another way to write Ek in terms of the ei that looks completely
different. In this other way, it turns out that Ek is expressed in terms of all the ei except ek and is given
by (only valid in Picture B where g=1)

Ek = det(R) (-1)k-1 e1 x e2 x ......x eN // ek missing k = 1,2,3...N (6.7.2)

This is a generalized cross product (see Section A.4) of N-1 vectors, since ek is missing, so there are N-2
"crosses". The above multi-cross-product equation is a shorthand for

(Ek)α ≡ det(R) (-1)k-1εαabc...x (e1)a(e2)b ...... (eN)x // κ and (eκ)K are missing (6.7.3)

Here ε is the totally antisymmetric tensor with N indices. If κ is the kth letter of the alphabet (k = 2 ⇒ κ =
b ), then κ is missing from the list of summation indices of ε, and the factor (ek)κ is missing from the
product of factors, so there are then N-1 factors.
This cross product expression for En is derived in Appendix A.
This is all fairly obscure sounding, but can be brought down to earth by writing things out for N = 3,
where the formula reduces to this cyclic set of equations,

E1 = det(R) e2 x e3
E2 = det(R) e3 x e1
E3 = det(R) e1 x e2 . (6.7.4)

These equations can be verified in a simple manner. To show an equation is true, if suffices to show that
the projections of both sides on the three en are the same, since the en form a complete basis as noted
earlier. For the first equation one needs then to show that

E1 • en = det(R) e2 x e3 • en for n = 1,2,3 . (6.7.5)

If n=2 or n=3, both sides vanish, according to en • Em = δn,m on the left, and according to geometry on
the right, which leaves just the case n = 1. In this case the LHS = 1, so one has to show that e2 x e3 • e1
= 1/det(R) = det(S). But

84
 Chapter 6: Reciprocal Base Vectors

e1 • e2 x e3 = (e1)i (e2 x e3)i = (e1)i εijk (e2)j(e3)k = εijk (e1)i(e2)j(e3)k

= εijk Si1Sj2Sk3 = det(S) QED // (3.2.5) says (en)i = Sin (6.7.6)

The other two equations of the set can be verified in the same manner.
Here is a picture (N=3) drawn in x-space for a non-orthogonal coordinate system. The vectors shown
here form a distorted right handed coordinate system which has det(R) > 0.

(6.7.7)

The reader is invited to exercise his or her right hand to confirm the directions of the arrows,

E1 = det(R) e2 x e3 E2 = det(R) e3 x e1 E3 = det(R) e1 x e2 . (6.7.8)

6.8 Comparison of ē n and En

Recall from (6.3.1) that ē n is the covariant partner of en where ē n = ḡ en .

Recall from (6.1.2) that En is the reciprocal of en where En ≡ g'ni ei .

The comparison of these two defining equations is interesting,

( ē n)i ≡ ḡ ik(en)k // matrix acts on vector index (6.8.1)

(En)i ≡ g'nk (ek)i . // matrix acts on ek label (6.8.2)

If x-space is Cartesian, then ē n = en as usual, but of course En ≠ en in this case since normally g' ≠1. We
mention this to head off a possible confusion when the Standard Notation is introduced in the next
Chapter and the above two equations become ( see (7.4.1) and (7.13.1) ),

(en)i ≡ gik(en)k // Standard Notation, g lowers an index (6.8.3)

(en)i ≡ g'nk (ek)i . // Standard Notation, k is a label on ek, not an index (6.8.4)

85
 Chapter 6: Reciprocal Base Vectors

The mapping to standard notation does not include ē n → en, for example. One fact about the standard
notation is that, unlike the developmental notation, one cannot look at a vector in bold like en and
determine whether it is contravariant or covariant. Only when the index is displayed can one tell. One can
think of en as representing both its contravariant self and its covariant partner (en is a tensorial vector).

6.9 Handedness of coordinate systems: the en , the sign of det(S), and Parity

1. Handedness of a Coordinate System. Let bn be a complete set of basis vectors in an N dimensional

vector space, where the bn are not necessarily of unit length, and are not necessarily orthogonal. Consider
this quantity

B ≡ det (b1, b2.....bn ) = εabc..x (b1)a (b2)b.... (bN)x = b1• [b2 x b3......x bN] (6.9.1)

where the generalized cross product is discussed in Section A.4. This basis bn defines a "coordinate
system" in that we can expand a position vector as follows

x = Σn x(b)n bn (6.9.2)

where the x(b)n are the "coordinates" of point x in this coordinate system. We make the following
definition:

system bn is a "right handed coordinate system" iff B>0

system bn is a "left handed coordinate system" iff B<0. (6.9.3)

One of course wants to show that for N = 3 this definition corresponds to one's intuition about right and
left handed systems. For N= 3 ,

B ≡ det (b1, b2, b3 ) = εabc (b1)a (b2)b(b3)c = b1• [b2 x b3] (6.9.4)

Suppose the bn are arranged as shown in this picture, where the visual intention is that the corner nearest
the label b1 is closest to the viewer.

(6.9.5)

With one's high-school-trained right hand, one can see that b2 x b3 points in the general direction of b1
(certainly b2 x b3 lies somewhere in the half space of the b2, b3 face plane which contains b1), and so the

86
 Chapter 6: Reciprocal Base Vectors

quantity B = b1• [b2 x b3] > 0. So this is an example of a right-handed coordinate system. The figure
shown is a skewed 3-piped which can be regarded as a distortion of an orthogonal 3-piped for which the
b would span an orthogonal coordinate system in which one would have ^
n b =^ b x^
1 b . 2 3

2. The x'-space e'n coordinate system is always right handed. In our standard picture of x-space and x'-
space, the coordinate system in x'-space is spanned by a set of basis vectors e'n where (e'n)i = δn,i , as
discussed in Chapter 3 (a). This system is "right handed" because

B = εabc..x (e'1)a (e'2)b.... (e'N) = εabc..x δ1aδ2b....δxN = ε123...N = +1 > 0 (6.9.6)

where we use the standard normalization of the ε tensor as shown. Notice that this conclusion is
independent of the metric tensor g' in x-space.

3. The x-space un coordinate system is always right handed. The basis un where (un)i = δn,i in x-space is
right handed for the same reason as shown in the above paragraph, and for any g. When g=1 in x-space,
the un form the usual Cartesian right-handed orthonormal basis in x-space. See Section 3.4 concerning the
meaning of "unit vector".

4. The x-space en coordinate system handedness is determined by the sign of det(S). Our x-space
coordinate system of great interest is that spanned by the en basis vectors, where (en)i = Sin as in (3.2.5).
One has

B = εabc..x (e1)a (e2)b.... (eN) = εabc..x Sa1 Sb2.... SxN = det(S) . (6.9.7)

Therefore, using σ ≡ sign(detS) and the Jacobian J = det(S) of (5.12.6), one has

system en is a "right handed coordinate system" iff det(S) = J > 0 or σ = +1

system en is a "left handed coordinate system" iff det(S) = J < 0 or σ = –1 . (6.9.8)

5. The Parity Transformation. The identity transformation F = 1 results in matrix SI = I with detSI = +1.
In this case the en form a right-handed coordinate system, and in fact en = un. The parity transformation
F = -1, on the other hand, results in SP = -I with det(SP) = (-1)N and en = -un. When N is odd, the parity
transformation converts the right-handed un system to a left-handed en system. If S is some matrix which
does not change handedness, meaning detS > 0, then S' = SSP does change handedness for odd N, since in
this case detS' = detS det SP = (-1)N detS. So given some S' that changes handedness for N=odd, one can
regard it as "containing the parity transformation" which, if removed, would result in no handedness
change. When N is even, handedness stays the same under F = -1.

6. N=3 Parity Inversion Example. Since x' = -x under the parity transform F = -1, parity is a reflection of
all position vectors through the origin. Objects sitting in x-space, such as N-pipeds, whether or not the
origin lies inside the object, are "turned inside out" by the parity transformation, but the inside of the
object still maps to the inside of the parity-transformed object under this transformation.

87
 Chapter 6: Reciprocal Base Vectors

Consider this crude picture which shows on the left a right-handed 3-piped in x-space where e1 and e2
happen to be perpendicular, and the back part of the 3-piped is not drawn. This 3-piped is associated with
some transformation S [ (en)i = Sin ] with detS > 0.

(6.9.9)

Now consider S' = SSP = SPS = -S. For this S', the 3-piped appears as shown on the right. The two pipeds
here are related by a parity transformation, all points inverting through the origin. On the right, the
volume of the 3-piped lies toward the viewer from the plane shown. The circled dot on the left represents
the out-facing normal vector of the 3-piped face area which is facing the viewer, and this normal is in the
direction – e1xe2. This is called a "near face" in Appendix B since it touches the tails of the en. After the
parity transformation, this same face has become the back face on the inverted 3-piped shown on the
right, with out-facing normal indicated by the X. The direction of this normal is + e1xe2 . In general, an
out-facing "near face" area points in the -En direction, and Appendix A shows that E3 = e1 x e2 / det(S).
On the left we have E3 = e1 x e2 / |det(S)| so the face there just mentioned points in the -E3 = – e1xe2
direction. On the right we have E3 = e1 x e2 / det(S') = - e1 x e2 /|det(S)|, so the face there points in the
-E3 = +e1xe2 direction.

7. The sign of det(S) in the curvilinear coordinates application. For a given ordering of the x' coordinates,
det(S) will have a certain sign. By changing the x'i ordering to any odd permutation of the original
ordering (for example, swap two coordinates), det(S) will negate because two columns of Sik(x') ≡
(∂xi/∂x'k) will be swapped. In the curvilinear coordinates application it is therefore always possible to
select the ordering of the x'i coordinates to cause det(S) to be positive. One always starts with a right-
handed Cartesian system ^ n for x-space, and det(S)>0 then guarantees that the en will form a right-handed
system there as well. Since the underlying transformation F is assumed invertible, one cannot have
det(S)=0 anywhere in the domain x (or range x') of x' = F(x), and therefore det(S) cannot change sign
anywhere in the space of interest.
For graphical reasons, we have selected coordinates in the "wrong order" in both the polar
coordinates examples (called Example 1) and in the elliptic polar system studied in Appendix C, which is
why detS < 0 for both these systems.

88
 Chapter 7: Standard Notation

7. Translation to the Standard Notation

In this Chapter we discuss the "translation" from our developmental notation (all lower indices; overbars
for covariant objects) to the Standard Notation used in tensor analysis.
The developmental notation has served well in the discussion of scalars and vectors, tensors of rank-0
and rank-1. For pure (unmixed) tensors of rank-2 it does especially well, allowing the use of matrix
algebra to leverage the use of familiar matrix theorems such as det(ABC) = det(A)det(B)det(C) and A-1 =
cof(AT)/det(A). The transformation of the contravariant metric tensor is cleanly expressed as g' = R g RT,
and so on. The notation in fact works fine for unmixed tensors of any rank, but runs into big trouble with
"mixed" tensors as shown in the next Sections.
We continue to use Picture A:

7.1 Outer Products

It is possible to form larger tensors from smaller ones using the "outer product" method. For example,
consider,

Tab ≡ UaVb (7.1.1)

where U and V are assumed to be contravariant vectors. One then has

T 'ab = U'aV'b = (Raa'Ua') (Rbb'Vb') = Raa' Rbb' Ua'Vb' = Raa' Rbb' Ta'b' (7.1.2)

so in this way a contravariant rank-2 tensor like M in (5.6.3) has been successfully constructed from two
contravariant vectors. Similarly,

T̄ab ≡ ŪaV̄b ⇒ T̄'ab = STaa' STbb' T̄a'b' (7.1.3)

so the outer product of two covariant vectors transforms as a covariant rank-2 tensor, again as in (5.6.3).

More generally, one can combine any number of contravariant tensors in an outer product fashion to
obtain a new tensor. Here are a few examples:

Tabc = UaAab

Tabcd ≡ AabBcd

Tabcde ≡ AabBcdUe (7.1.4)

89
 Chapter 7: Standard Notation

7.2 Mixed Tensors and Notation Issues

Suppose we take the "outer product" of a contravariant vector with a covariant vector,

[ ... ]ab ≡ UaV̄b (7.2.1)

where we are not sure what to call this thing, so we just call it [...]. Here is how this new object transforms
(always: with respect to the underlying transformation x' = F(x) )

[ ... ]'ab = U'aV̄'b = (Raa'Ua') (STbb'V̄b') = Raa' STbb' Ua'V̄b' = Raa' STbb' [...]ab . (7.2.2)

This object transforms as a contravariant vector on the first index (ignoring the second), and as a
covariant vector on the second index (ignoring the first). This is an example of a "mixed" rank-2 tensor.
Extending this outer product idea, one can make elaborate tensor objects with an arbitrary mixture of
"contravariant indices" and "covariant indices". For example

[.....]abcd = UaV̄b XcȲd . (7.2.3)

To write down the transformation rule for such an object, one must know which indices are contravariant
and which are covariant. It is totally clear how the object transforms, looking at the right side of (7.2.3),
but somehow this information has to be embedded in the notation [.....]abcd because once this object is
defined, the right hand side might not be immediately available for inspection. Worse, there may be no
right hand side for a mixed tensor, because not all mixed tensors are outer products of vectors (they just
transform as if they were).
Just as we can use the idea V̄ ≡ ḡ V to convert a contravariant vector to its covariant partner, we can
similarly use ḡ to convert the 1st or 3rd index on [.....]abcd from contravariant to covariant. We could
apply two ḡ's with the proper linkage of indices to convert them both at once.
So given the ability of ḡ to change any index one way, and g to change it the other way, one can think
of the 4-index object [.....]abcd as a family of 16 different 4-index objects, each corresponding to a certain
choice for the indices being one type or the other. We know how to interconvert between these 16 objects
just applying g or ḡ factors.
So how does one annotate which of the 16 objects [.....] one is staring at for some choice of index
types? Here is a somewhat facetious possibility, the Morse Code method

–
Wab ≡ UaV̄b
--
W abcd = UaV̄b XcȲd . (7.2.4)

Instead of having a bar over the entire object, in the first case the bar it is placed just over the right side of
the W to indicate that b is a covariant index, while no bar means the first index is contravariant. The
second example shows how horrible such a notation would be.
We really want to put some kind of notation on the individual indices, not on the object! Here is a
notation that is slightly better than the Morse code option, though similar to it,

Wab-cd- = UaV-b XcY-d (7.2.5)

90
 Chapter 7: Standard Notation

Here overbars on covariant indices distinguish them. Now one can dispense with the overbars on
covariant vectors as well, putting the overbar on the index, for example V̄a = ḡ abVb → V-a = g -a-bVb .
There are several problems with this scheme. One is that in the spinor application of tensor analysis
used in special relativity, dots are placed on certain indices, such as in (5.14.8), and these would conflict
with the proposed overbars. A more substantial reason is that this last notation is hard to type (or typeset,
as one used to say), it looks cluttered with all the overbars, and the subscripts are already hard to read
without extra decorations since they are in a smaller font than the main text.

7.3 The up/down bell goes off

This is where a bell went off somewhere, perhaps in the mind of Gregorio Ricci in the 1880-1900 time
frame (1900 snippet quoted in (7.10.8) below). Someone might have said: suppose, instead of using
overbars on indices or some other decoration, we distinguish covariant indices by making them be
superscripts instead of subscripts. Superscripts are as easy to type as subscripts, and the result is fairly
easy to read and totally unambiguous. We would then have for our ongoing example of (7.2.4 and 5),

Wabcd = UaVbXcYd // a path not taken

This is almost what happened, but the up/down decision went the other way and we now have:

superscripts = contravariant = up
subscript = covariant = down (7.3.1)

and then we get this translation for (7.2.4),

Wab-cd- = UaV-b XcY-d → Wabcd = UaVbXcYd // the path taken (7.3.2)

and this has become The Standard Notation. Perhaps the reason for this choice was that the covariant
gradient ∂n object appeared more commonly in equations than idealized objects such as dx, and ∂n
already used a lower index.
A downside of this particular up/down decision is that every student has be be confused by the fact
that his or her familiar position, velocity and momentum vectors that always had subscripts suddenly have
superscripts in the Standard Notation. The silver lining is that this shocking change alerts the student to
the fact that whatever subject is being studied is going to have two kinds of vectors.
Despite appearances, it is not completely obvious how one should translate the whole world as
presented in the previous six Chapters into this new notation. There are quite a few subtle details that will
be discussed in the following Sections.

91
 Chapter 7: Standard Notation

7.4 Some Preliminary Translations; raising and lowering tensor indices with g

In the rest of this entire Chapter, anything to the left of a → arrow is in "developmental notation", while
anything to the right of → is in "Standard Notation".

1. Basic translations.

So we start translating some of the results above:

s→s // a scalar

Va → Va // a contravariant rank-1 tensor (vector)

V̄a → Va // a covariant rank-1 tensor (vector)

Mab → Mab // a contravariant rank-2 tensor

M̄ab → Mab // a covariant rank-2 tensor

gab → gab // the contravariant rank-2 metric tensor g (and similarly for g')
ḡ ab → gab // the covariant rank-2 metric tensor ḡ
g is inverse of ḡ → gab is inverse of gab . (7.4.1)

As noted earlier, one "feature" of the Standard Notation is that it is no longer sufficient to specify an
object by a single letter. One has to somehow indicate the index nature by showing index positions. Thus,
"g" stands for all four metric tensors gab , gab, gab and gab. The pure covariant metric tensor is gab or
perhaps g** . At first this seems a disadvantage of the notation, but one then realizes that the true object
really is "g", and it has four different "representations" and the notation makes this very clear.
In this same spirit, we have the following translation for a bolded vector:

V → V
V̄ → V. (7.4.2)

The reason is that the overbar is no longer used to denote covariancy. The above lines show a subtle
change in the interpretation of the bolded symbol V in the standard notation: the single symbol V stands
for both the developmental vector V and for its developmental covariant partner vector V̄. The new
symbol V is both contravariant with components Vn and it is covariant with components Vn.
(7.4.3)
2. Raising and Lowering Tensor Indices with g

As for converting a vector from one type to the other, the relations (5.8.4) become,

V̄a = ḡ abVb → Va = gabVb // gab "lowers" a contravariant index

Va = gab V̄b → Va = gab Vb // gab "raises" a covariant index (7.4.4)

In the Standard Notation, the nature of a vector (contravariant or covariant) is determined by the up or
down position of the index. Looking at gabVb = Va, we see that " gab lowers the index on Vb to give Va "

92
 Chapter 7: Standard Notation

which of course just means that ḡ abVb = V̄a in the developmental notation. It means nothing more, and
nothing less. Similarly, gab Vb = Va says that " gab raises the index on Vb to give Va " and this is just
the Standard Notation way of saying gab V̄b = Va.
The key idea here is that gab lowers a tensor index of a tensor, and gab raises a tensor index (both
with an implied sum as above). We know from the outer product idea shown in (7.1.1) that we can
construct a covariant rank-2 tensor from two covariant rank-1 tensors in this way.

Tab ≡ UaVb (7.1.1)

This is in Standard Notation now, so unlike (7.1.1), all three objects are covariant tensors. We can then
use gca to raise the a index on both sides of this equation,

gcaTab = Tcb // implied sum on a -- always implied sums on repeated indices!

gca(UaVb) = ( gcaUa)Vb = UcVb ⇒

Tcb = UcVb . (7.4.5)

Here, Tca is a "mixed" rank-2 tensor, the first index is contravariant, the second index is covariant. This
fact matches the nature of the right side of the equation UcVb. We could have but did not deal with mixed
tensors in the developmental notation because the notation could not really handle it well, as shown in
(7.2.4). So in developmental notation, we dealt only with "pure" rank-2 tensors like Mab and M̄ab.

As a next step, we can apply gdb to both sides of (7.4.5),

gdb Tcb = Tcd

gdb (UcVb) = Uc(gdbVb) = UcVd ⇒

Tcd = UcVd (7.4.6)

We end up with a contravariant rank-2 tensor that is the outer product of two contravariant rank-1 tensors.
This is in fact the equation (7.1.1) expressed in Standard Notation.

Now let's do both index raising operations at the same time:

Tab ≡ UaVb

gca gdb Tab = Tcd

gca gdb UaVb = ( gcaUa)(gdbVb) = UcVd ⇒

Tcd = UcVd (7.4.7)

93
 Chapter 7: Standard Notation

This shows that we are free to raise or lower corresponding indices on both sides of a tensor equation at
will, and we don't have to write out the details. Thus if the first of these is true, then the other three forms
are also valid:

Tab ≡ UaVb ⇒ Tab = UaVb Tab = UaVb Tab = UaVb (7.4.8)

This idea works with any rank-2 tensor because any rank-2 tensor transforms under x' = F(x) in the same
way that the outer product of two vectors transforms. If M is some arbitrary contravariant rank-2 tensor,
then gcaMab = Mcb and so on, repeating all of the above.
The idea applies as well to rank-3 tensors and rank-n tensors as we shall describe below in Section
7.10. For example, if Tabc = UaVbWc, or if M is some arbitrary rank-3 tensor Mabc, then

Tabc ≡ UaVbWc Mabc

gdaTabc = Tdbc gdaMabc = Mdbc

(gdaUa)VbWc = UdVbWc ⇒

Tdbc = UdVbWc . (7.4.9)

In this case, the validity of the first equation implies the validity of the 7 others:

Tabc ≡ UaVbWc ⇒ Tabc ≡ UaVbWc Tabc ≡ UaVbWc Tabc ≡ UaVbWc

Tabc ≡ UaVbWc Tabc ≡ UaVbWc Tabc ≡ UaVbWc

and finally Tabc ≡ UaVbWc (7.4.10)

To summarize, in this new Standard Notation, the covariant metric tensor gab becomes an "index
lowering operator" and the contravariant metric tensor gab becomes an "index raising operator". This is a
huge advantage of the Standard Notation. It pretty much eliminates the need to think, something
universally appreciated. In a certain obscure sense, it is like double entry accounting (credits and debits),
where the notation itself serves as a check on the accuracy of bookkeeping entries.

The g raising and lower rules may be represented generically in this manner,

gaa' [----a'---] = [----a---]

gaa' [----a'---] = [----a---] (7.4.11)

where [----a'---] represents an arbitrary tensor expression (perhaps just one tensor) with lots of tensor
indices indicated by dashes. Each of these dash indices could be up or down, it does not matter. An
example of the second line is gaa'Ma'bc = Mabc .

94
 Chapter 7: Standard Notation

3. The invariant distance and covariant dot product:

Based on the discussion so far, we can write the following translations, starting from (5.2.5) and (5.10.3),

dxi → dxi

(ds)2 = ḡ ab dxa dxb → gabdxadxb

A•B = ḡ ab Aa Bb → A•B = gab Aa Bb . (7.4.12)

Using the raising and lowering idea above, we can write the dot product in other ways:

A•B = gab Aa Bb = Aa( gabBb) = AaBa

A•B = gab Aa Bb = gba Aa Bb = (gba Aa)Bb = AbBb // g is symmetric

A•B = AbBb = Ab (gbcBc) = gbcAbBc . (7.4.13)

To summarize :

A•B = gab Aa Bb = AaBa = AaBa = gabAaBb = B •A . (7.4.14)

The general idea is this: any tensor index on any tensor object can be raised by gab and can be lowered by
gab. Remember that a tensor object lives in some space like x-space, so we shall have to ponder what to
do for our matrices Sab and Rab which live half in x-space and half in x'-space, a subject we defer to
Section 7.5.

4. Contraction Rules

When an index is summed with one index down and the other up, one says that the two indices are
contracted. For example, in the expressions AaBa or AaBa, index a is contracted. In gabAaBb both
indices a and b are contracted.

Tilt Reversal Theorem. The tilt of any pair of contracted indices can be reversed without affecting
anything. (7.4.15)

A proof is given below in Section 7.11. An example is AaBa =AaBa as shown above in (7.4.14).

Contraction Theorem. In analyzing the transformation nature of a tensor expression, one can simply
ignore all contracted indices and look only at remaining indices to determine the tensor nature of the
object. (7.4.16)

A proof is given below in Section 7.12. As examples, in the cases AaBa or gabAaBb, if we ignore
contracted indices, there are no indices left, so these objects must transform as a tensor with no indices,

95
 Chapter 7: Standard Notation

which is a tensorial scalar. Thus A•B above is a scalar, something we already knew from (5.10.2). The
object AaMab transforms as a covariant vector, and the object Maa transforms as a scalar.

5. What about having g raise and lower indices on itself?

After all, g is a valid tensor. We find

gabgbc = gac // lower the first index of rank-2 tensor gbc

gacgbc = gba // lower the second index

gcbgba = gca . // lower both indices (lower second, then lower first) (7.4.17)

We know from (7.4.1) above that gab and gab are inverses. Thus, it must be true that

gabgbc = (1)ac = δac = δa,c

gbcgac = (1)ba = δba = δb,a . (7.4.18)

Here δac and δba are just cosmetically nice ways to write the Kronecker deltas shown. Comparing
(7.4.17) and (7.4.18) we learn that

gac(x) = δac = δa,c

gba(x) = δba = δb,a . (7.4.19)

Thus, the mixed forms of the metric tensor are trivial and are not functions of the nature of x-space.

7.5 Dealing with the matrices R and S ; various Rules and Theorems

1. Translations of R and S

Consider the translation of this partial derivative into the new up/down notation. Since the differential dx
element is contravariant and is now written dxi :

(∂x'i/∂xk) → (∂x'i/∂xk) . (7.5.1)

In terms of "existence", this object has one leg in each space of Picture A of (1.11). The gradient operator
∂/∂xk is an x-space object, while x'i is an x'-space object. Since (∂x'i/∂xk) does not live in x-space or in
x'-space exclusively, but straddles the two spaces, it cannot possibly be a tensor of any kind. Recall that a
tensor object must be entirely within a space, it cannot have body parts hanging out into other spaces.
Nevertheless, it seems clear that each of the two indices has a well-defined nature. We showed in (2.4.2)
that the gradient transforms as a covariant vector, so we regard k as a covariant index and write ∂/∂xk =

96
 Chapter 7: Standard Notation

∂k. And of course dx'i is a contravariant vector, so i is a contravariant index. Here then is the proper
translation starting with (2.5.1),

Rik ≡ (∂x'i/∂xk) → Rik ≡ (∂x'i/∂xk) = ∂kx'i . (7.5.2)

To summarize, Rik is not a mixed rank-2 tensor, though it looks just like one. Therefore, Rik can never
appear in a tensor equation -- it just appears in the equations that show how tensors transform. However,
each of the two indices of R has a well-defined transformational nature, and we place them up and down
in the proper manner.
It is not atypical for an object to have up and down indices but the object is not a tensor. As was noted
below (2.3.2), the canonical example is that for a non-linear transformation x' = F(x), xi has a
contravariant index but is not (does not transform under F as) a contravariant vector.
Consider now the translation of the transformation rule for a contravariant vector from (2.5.1),

V'a = RabVb → V'a = RabVb . // contravariant (7.5.3)

Even though R is not a tensor, we see that index b is contracted and is thus neutralized from the
evaluation of the tensor nature of the RHS. This leaves upper index a as the only free index, indicating
that the RHS is a contravariant vector, and this of course then matches the LHS. However, it is a
contravariant vector in x'-space, which will be further discussed in (7.5.9) below. Main point: we can
deal with the indices on R just as we deal with indices on true tensors.
Notice that, even though both sides of V'a = RabVb have the same "tensor nature" (both sides are a
contravariant vector component in x'-space) one cannot ask how the equation V'a = RabVb "transforms"
under a transformation. That question can only be asked about equations constructed of objects all of
which are tensors in the same space. Here V and half of R are in x-space, and V' and the other half of R
are in a x'-space. There is no object called R', as if R were in x-space and R' were in x'-space.

We can repeat the above discussion for S instead of R. We omit the words and just show the translations,
quoting expressions from (2.1.6) and (2.5.1) :

(∂xi/∂x'k) → (∂xi/∂x'k)

Sik ≡ (∂xi/∂x'k) → Sik ≡ (∂xi/∂x'k) = ∂'kxi

V̄'a = STabV̄b = Sba V̄b → V'a = SbaVb // covariant (7.5.4)

2, Transformation rule for rank-2 tensors

For a contravariant rank-2 tensor, the translation of the transformation rules (5.6.3) are,

M'ab = Raa'Rbb'Ma'b' → M'ab = Raa'Rbb'Ma'b' contravariant rank-2 tensor

(7.5.5)
M̄'ab = Sa'aSb'bM̄a'b' → M'ab = Sa'aSb'bMa'b' . covariant rank-2 tensor

Applying this to the metric tensor M = g we find these translations of (5.7.6),

97
 Chapter 7: Standard Notation

g' = R g RT ⇒ g'ab= Raa'Rbb'ga'b' → g'ab = Raa'Rbb'ga'b'

(7.5.6)
ḡ' = ST ḡ S ⇒ ḡ'ab = Sa'aSb'b ḡ a'b' → g'ab = Sa'aSb'b ga'b' .

The inverses of the above equations appear in (5.7.7) and translate this way ,

g = S g' ST ⇒ gab = Saa'Sbb'g'a'b → gab = Saa'Sbb'g'a'b

(7.5.7)
T a' b'
ḡ = R ḡ' R ⇒ ḡ ab = Ra'aRb'b ḡ'a'b' → gab = R aR b g'a'b' .

By raising and lowering indices on the first result in (7.5.5), and using the tilt reversal rule (7.4.15), we
obtain these four transformation rules :

M'ab = Raa' Rbb' Ma'b'

M'ab = Raa' Rbb' Ma'b'
M'ab = Raa' Rbb' Ma'b'
M'ab = Raa' Rbb' Ma'b' . (7.5.8)

One sees then a family of four tensors associated with M. The first is contravariant, the last is covariant,
and the other two are mixed. Comparison of the last line of (7.5.8) to the last line of (7.5.5) suggests that
Raa' = Sa'a , and we will formally prove this fact in (7.5.13) below.

3. Raising and lowering indices on R and S

Although R and S are not tensors, one can still raise and lower their two indices using metric tensors, but
things are a little different from the tensor situation, the reason being that R and S each have one foot in x-
space and the other foot in x'-space.

Object Rab = (∂x'a/∂xb) was considered above. One could lower the a index using g'** since x'a is in x'-
space and is an up index. The index in ∂/∂xb = ∂b is really a lower index (gradient), so one could in effect
raise it using g** (no prime) because ∂/∂xb is in x-space. So when raising and lowering indices on Rab
one has the unusual situation that one must use g' when acting on the first index, and g when acting on the
second. With this in mind, we can now write three other index configurations of Rab :

Rab = (∂x'a/∂xb) // original object (formerly Rab)

Rab = Rab' gb'b = (∂x'a/∂xb) // g pulls up the second index of Rab
Rab = g'aa'Ra'b = (∂x'a/∂xb) // g' pulls down the first index of Rab
Rab = g'aa'Ra'b' gb'b = (∂x'a/∂xb) . // both actions at once (7.5.9)

Although the g and g' factors can be placed anywhere, we have put g' factors on the left of R, and g
factors on the right, each next to its appropriate leg of R.

98
 Chapter 7: Standard Notation

In each case, examination of the corresponding partial derivative shows that that the index sense matches
on both sides. For example, in Rab = (∂x'a/∂xb) = ∂bx'a, both indices are contravariant on both sides.
Remember that Rab is not a contravariant rank-2 tensor due to its dual-space nature.

In the same manner, we arrive at these index configurations for Sab :

Sab(x) = (∂xa/∂x'b) // original object (formerly Sab)

Sab ≡ Sab' g'b'b = (∂xa/∂x'b) // g' pulls the second index up
Sab ≡ gaa'Sa'b = (∂xa/∂x'b) // g pulls the first index down
Sab ≡ gaa'Sa'b' g' b'b = (∂xa/∂x'b) // both actions at once (7.5.10)

For object S, the metric tensor g raises or lowers the first index of S, while g' raises or lowers the second
index of S. This is just the reverse of what happens for object R as reported above. This is not surprising
since one gets S↔R when one swaps x-space ↔ x'-space.

4. Inverse of R and S (and Comment on Transpose)

We have showed the translation Rab → Rab . In the standard notation, imagine that there is some inverse
R-1 defined by (R-1)caRab = δcb. The chain rule says that

(∂xc/∂x'a) (∂x'a/∂xb) = δcb or Sca Rab = δcb

(∂x'c/∂xa) (∂xa/∂x'b) = δcb or Rca Sab = δcb . (7.5.11)

Therefore from the first line it must be that (R-1)ca = Sca. The second line shows that (S-1)ca = Rca.
Using the above rules for raising and lowering indices on both sides of an equation, we obtain these four
versions of the developmental notation fact that R-1 = S and S-1 = R :

(R-1)ik = Sik (S-1)ik = Rik

(R-1)ik = Sik (S-1)ik = Rik
(R-1)ik = Sik (S-1)ik = Rik
(R-1)ik = Sik (S-1)ik = Rik . (7.5.12)

In (7.5.11) we see a certain "up tilt" matrix multiplication which will be explained below in Section 7.8.

99
 Chapter 7: Standard Notation

5. The R-S Tilt Theorem: Sab = Rba which is the same as: (∂xa/∂x'b) = (∂x'b/∂xa) (7.5.13)

Notice in this rule that the indices are reflected in a vertical line running between the indices: Sa|b = Rb|a.

Proof: This proof is a bit long-winded, but brings in many earlier results and is a good exercise:

δba" Saa" = Sab // introduce a δ . Remember all g's are symmetric.

(g' bb' g' a"b') Saa" = Sab // since g'ab and g'ab are inverses, see (7.4.18)

g' bb' δb'b" Saa" g' a"b"= Sab // reorder and use g' a"b' = δb'b"g' a"b"

g' bb' (Rb'a' Sa'b") Saa" g'a"b"= Sab // δb'b" = (Rb'a' Sa'b") from (7.5.11)

g' bb' Rb'a' (Saa" Sa'b" g'a"b") = Sab // regroup (7.5.14)

g' bb' Rb'a' (gaa') = Sab // gaa' = Saa" Sa'b" g'a"b" from (7.5.7)

(g' bb' Rb'a' gaa') = Sab // regroup

Rba = Sab . // use last line of (7.5.9)

Notice that the above theorem says

Sab = (∂xa/∂x'b) = (∂x'b/∂xa) = Rba . (7.5.15)

Similar results can be derived for other index positions (or we can just raise and lower indices!) to get

Sab = Rba = (∂xa/∂x'b) = (∂x'b/∂xa)

Sab = Rba = (∂xa/∂x'b) = (∂x'b/∂xa)
Sab = Rba = (∂xa/∂x'b) = (∂x'b/∂xa)
Sab = Rba = (∂xa/∂x'b) = (∂x'b/∂xa) . (7.5.16)

Here index a is always in x-space, while index b is in x'-space.

6. Determinants of R and S.

We showed earlier that Rij→ Rij and Sij→ Sij. The determinants det(R) and det(S) translate as
follows:

det(R) = εabc...R1aR2b....RNx → det(R**) = εabc...R1aR2b....RNx

det(S) = εabc...S1aS2b.....SNx → det(S**) = εabc...S1aS2b....SNx
or

100
 Chapter 7: Standard Notation

det(R) = εabc...Ra1Rb2....RxN → det(R**) = εabc...Ra1Rb2....RxN

det(S) = εabc...Sa1Sb2.....SxN → det(S**) = εabc...Sa1Sb2....SxN (7.5.17)

where ε is the bookkeeping permutation tensor discussed in Section 7.7 below.

Reader Exercise: Write det(R**) = (εabcR1aR2bR3c) and det(S**) = (εa'b'c'Sa'1Sb'2Sb'3 ). We know

that in either developmental or standard notation det(R)det(S) = det(RS) = det(1) = 1. The exercise is to
verify this directly using the N=3 first expression in (D.10.37) for εabcεa'b'c' along with (7.5.11).

7. Jacobian and related

Here are translations of a few equations from Section 5.12:

(5.12.2) det(S) = 1/det(R) → det(Sij) = 1/ det(Rij) (7.5.18)

(5.12.6) J = det(S) → J = det(Sij) (7.5.19)

(5.12.12) g = det(ḡ) → g = det(gij)

g' = det(ḡ') → g' = det(g'ij) (7.5.20)

(5.12.14) 1/g = det(g) → 1/g = det(gij)

1/g' = det(g') → 1/g' = det(g'ij) (7.5.21)

(5.12.14) g' = J2 g → g' = J2 g

|J| = g'/g → |J| = g'/g (7.5.22)

7.6 Orthogonality Rules, Inversion Rules, Cancellation Rules

1. Orthogonality Rules:

Theorem (7.5.13) that Sab = Rba can be used to eliminate S in various forms of SR = 1:

SR = 1 → Sab Rbc = δac ⇒ Rba Rbc = δac ⇒ Rba Rbc = gac from (7.4.19) . (7.6.1)

Although R is not a tensor, the object Rba Rbc is a tensor of the mixed type Mac. Thus, both sides of this
last equation transform as this type of mixed rank-2 tensor.

In this last equation, we can lower index a on both sides, raise index c on both sides, and reverse the tilt of
the b contraction to get this result, again using (7.4.19),

Rba Rbc = gac = δac . (7.6.2)

We now repeat this process starting instead with RS = 1:

101
 Chapter 7: Standard Notation

RS = 1 → Rab Sbc = δac ⇒ Rab Rcb = δac ⇒ Rab Rcb = gac

Rab Rcb = gac = δac . (7.6.3)

Collecting the four forms written with δ we obtain the four orthogonality rules for R:

1: Rba Rbc = δac 2: Rba Rbc = δac // sum is on first index

3: Rab Rcb = δac 4: Rab Rcb = δac // sum is on second index (7.6.4)

For any general transformation x' = F(x), these rules are valid. Going the reverse direction, if we replace
all Rxy with Syx, the four orthogonality rules of (7.6.4) can be rederived in this manner,

1: Rba Rbc = Rba Scb = ScbRba = (SR)ca = (1)ca = δca = δac // down-tilt

2: Rba Rbc = RbaScb = ScbRba= (SR)ca = (1)ca = δca = δac // up-tilt

3: Rab Rcb = Rab Sbc = (RS)ac = (1)ac = δac // up-tilt

4: Rab Rcb = RabSbc = (RS)ac = (a)ac = δac // down-tilt

All these rules just say RS = SR = 1. Notice in the last four lines the notion of up-tilt and down-tilt matrix
multiplication. This subject is discussed in Section 7.8 below.

Comment: The orthogonality rules are easy to memorize using these few facts:
• The sum is either on both first indices, or on both second indices
• the tilts of the two R's are opposite each other.

2. Inversion Rules.

The first rule makes this claim:

[----a-----] = Rab [--------b------] ⇔ Rab [----a-----] = [--------b------] (7.6.5)

sum on 2nd index b sum on 1st index a

so if one wants to invert the left equation for [--------b------], the result is as shown on the right. Notice
that the R indices are reflected through a horizontal line segment between the indices. This is different
from the fact (7.5.13) that Sab = Rba where indices are reflected through a vertical line segment between
the indices.

Proof: Start with [----a-----] = Rab' [--------b'------]. Apply Rab to both sides, then use (7.6.4) :

Rab[----a-----] = Rab Rab' [--------b'------] = δbb' [--------b'------] = [--------b------] QED

102
 Chapter 7: Standard Notation

The rule is also valid if all tilts are reversed. Thus,

[----a-----] = Rab [--------b------] ⇔ Rab [----a-----] = [--------b------] (7.6.6)

sum on 2nd index b sum on 1st index a

Proof: Start with [----a-----] = Rab' [--------b'------]. Apply Rab to both sides, then use (7.6.4) :

Rab[----a-----] = Rab Rab' [--------b'------] = δbb] [--------b'------] = [--------b------] QED

Example of (7.6.5): V'a = RabVb ⇔ RabV'a = Vb or Vb = RabV'a (7.6.7)

Example of (7.6.6): V'a = RabVb ⇔ RabV'a = Vb or Vb = RabV'a (7.6.8)

Example: Rank-3 tensor transformation inversion :

M'abc = RaiRbjRckMijk // rank-3 transformation rule (contravariant)

Mabc = RiaRjbRkcM'ijk // its inverse (7.6.9)

Notice that, in the inverse transformation, the R indices are up-tilted and the summation indices are
strangely on the left side. Using the raising and lowering rules of (7.5.9), one sees that the corresponding
covariant forms are obtained just by moving indices vertically on both sides. Thus.

M'abc = RaiRbjRckMijk // rank-3 transformation rule (covariant)

Mabc = RiaRjbRkcM'ijk // its inverse (7.6.10)

3. Cancellation Rules

The first rule makes this claim:

Rab [----b-----] = Rab [--------b------] ⇔ [----b-----] = [--------b------] (7.6.11)

where the two R factors can be cancelled out.

Proof: Start with Rab' [----b'-----] = Rab' [--------b'------]. Apply Rab to both sides, then use (7.6.4) :

RabRab' [----b'-----] = RabRab' [--------b'------] ⇒

δbb' [----b'-----] = δbb' [--------b'------] ⇒ [----b-----] = [--------b------] QED

The rule is also valid if all tilts are reversed. Thus,

Rab [----b-----] = Rab [--------b------] ⇔ [----b-----] = [--------b------] (7.6.12)

The proof is similar to that shown above.

103
 Chapter 7: Standard Notation

7.7 About δ and ε

The Kronecker δ is sometimes written in different ways to make things "look nice",

δab = δab = δba = δba = δa,b . (7.7.1)

Eq (7.4.19) showed that that one can regard the above sequence of equalities as saying

gab = gab = gba = gba = δa,b (7.7.2)

where these g objects are mixed versions of the symmetric rank-2 metric tensor gab. There is no "δ
tensor", it is the g metric tensor, but tradition is to write the diagonal objects using the δ symbol.

The object εabc... is a bit more complicated. It can at first be regarded as a mere bookkeeping device, in
which context it is usually called "the permutation tensor". It appears for example in the expansion of a
determinant in N dimensions,

det(M) = εabc...xM1aM2b.....MNx = εabc...xMa1Mb2.....MxN (7.7.3)

or in an ordinary cross product in Cartesian space,

Aa = εabcBbCc . (7.7.4)

This permutation tensor has the usual properties that ε123...N = +1, that ε changes sign when any two
indices are swapped, and that ε vanishes if two or more indices are the same. This permutation "tensor" is
not really a tensor since one would regard it as being the same in x-space or x'-space. Whether indices are
written up or down on this ε is immaterial.
At another level, however, εabc...x with N indices (the same ε symbol is used) is a covariant rank-N
tensor density of weight -1 known as the Levi-Civita tensor. This subject is addressed in Appendix D in
much detail. In what we call the Weinberg convention, individual indices of ε can be raised and lowered
by g as discussed in Section 7.4 subsection 2, just as with any tensor. Therefore, in Cartesian space with g
= 1, indices on ε are raised and lowered with no consequence [ see (5.9.1) or (7.4.4) ], and then one can
identify any form of ε as being the permutation tensor. For example, εabc = εabc = εabc and so on. In a
non-Cartesian x-space, however, one would say that εabc = gbb'εab'c ≠ εabc. In the Weinberg
convention, one sets ε123..N = ε'123..N = 1 and εabc..x = ε'abc..x has the properties of the permutation
tensor described above and these properties are the same in x-space as in x'-space. Then for general g≠1,
εabc..x (lower indices) is NOT the permutation tensor. The bottom line is that one must be aware of the
space in which one is working (the Picture). The ε appearing above in the determinant expansion (7.7.3)
is always just the permutation tensor, but in the cross product that is not the case, and one would properly
write

Aa = εabcBbCc (7.7.5)

104
 Chapter 7: Standard Notation

and conclude that the cross product of two ordinary contravariant vectors is a covariant vector density
(Section D.8). Again, in Cartesian space where one often works, this would be the same as Aa =
εabcBbCc = εabcBbCc , but the "properly tilted form" Aa = εabcBbCc reveals the tensor nature of the
object Aa. As mentioned below in Section 7.15, this "covariant" equation would appear as A'a =
ε'abcB'bC'c in x'-space, but since A'a is a covariant vector density, A'a ≠ RabAb, and in fact A'a = J
RabAb.
The permutation tensor εabc...x and the contravariant Levi-Civita tensor εabc...x are both "totally
antisymmetric" which just means ε changes sign if any pair of indices is swapped. In fact, as discussed in
Section D.3, there IS only one antisymmetric tensor of rank N apart from a multiplicative scalar factor,
and εabc...x is it. This fact simplifies various calculations. Technically, εabc...x is a totally
antisymmetric tensor density, but normally it is just called "the totally antisymmetric tensor". As implied
by (D.5.1), the covariant Levi-Civita tensor εabc...x is also totally antisymmetric and is therefore a
multiple of εabc...x.
The reader is invited to peruse Appendix D at some appropriate time for more about tensor densities
and the ε tensor.

7.8 Covariance and Matrix Multiplication

Before continuing the process of translation from developmental to standard notation, we digress
momentarily to consider the notion of covariance in developmental notation.

As we shall discuss in more detail below in Section 7.15, an equation is said to be covariant under the
transformation x' = F(x) if it has "the same form" in both x-space and x'-space. The "same form" means
that the equation looks the same but everything is primed in x'-space.

Example 1: Newton's Law F = ma is covariant under rotations (x' = F(x) = Rx), and in x'-space this law
takes the form F' = m'a' which has the same form as the equation in x-space F = ma. Once we know that
F and a are contravariant vectors and m is a scalar, this conclusion is automatic from (2.3.2),

F = ma ⇒ F' = m'a' proof: F' = RF = R(ma) = m Ra = m a' = m' a' (7.8.1)

where m = m' follows since mass is a scalar under rotation. Thus, Newton's Law has the same form when
it is examined in two frames of reference related by a rotation. It is covariant under rotations.

Example 2: Consider the equation A•B = π where A and B are contravariant vectors and • is the covariant
dot product shown in (5.10.3), A•B ≡ ḡ abAaBb. It was shown in (5.10.2) that the quantity A•B
transforms as a scalar under general transformation x' = F(x) so that A'•B' = A•B. Since the number π is
also a scalar under any transformation (it is a constant), one could say that π' = π (it is the same number
3.14159 in x'-space and x-space), so

A•B = π ⇒ A'•B' = π' , equation is covariant. (7.8.2)

What we see here is that an equation is covariant IFF both sides of the equation transform as the same
tensorial tensor type under the transformation of interest. In Example 1, both sides of F = ma transform as

105
 Chapter 7: Standard Notation

contravariant vectors under rotations, and in Example 2 both sides of A•B = π transform as scalars under
a general transformation.

Example 3: Consider the outer product equation (7.1.1) Tab = UaVb where U and V are contravariant
vectors. We show in (7.1.2) that Tab transforms as a contravariant rank-2 tensor. Both sides of this
equation transform in this way, so in x'-space the equation becomes T'ab = U'aV'b. The equation is
therefore covariant under the transformation x' = F(x) with dx' = Rdx.
Approaching this example in a slightly different manner, suppose we define Tab ≡ UaVb where U and
V are contravariant vectors. We then ask: Is Tab a contravariant rank-2 tensor? Line (7.1.2) shows that
the answer is yes,

T 'ab = U'aV'b = (Raa'Ua') (Rbb'Vb') = Raa' Rbb' Ua'Vb' = Raa' Rbb' Ta'b' (7.8.3)

which matches the transformation rule as stated in (5.6.3).

Example 4: Suppose A and B are tensorial contravariant rank-2 tensors. Is the equation AB = C
covariant? If it were, we would have to show that in x'-space we have A'B' = C' where C is a contravariant
rank-2 tensor. To investigate, we use the rule (5.7.1) which states how a contravariant rank-2 tensor
transforms in terms of Picture A shown in (5.7.2) :

A'B' = (RART)(RBRT) = RA(RTR)BRT // since A and B are contra rank-2 tensors (7.8.4)

C' = RCRT = RABRT // assuming C is also a contra rank-2 tensor and AB = C

If it were true that RTR = 1, one would find from the first line above that A'B' = RABRT = RCRT = C' and
the answer would be yes, the equation AB = C is covariant. However, for a general Picture A
transformation with metric tensor g in x-space and g' in x'-space, what we know about R comes from
(5.7.6) : g' = R g RT . Even if g = 1 so x-space is Cartesian, this says g' = RRT, but this tells us nothing
about RTR. So for a general transformation, we have RTR ≠ 1 and so the equation AB = C is NOT
covariant. [In the special case that R is a rotation, so RT = R-1 (real orthogonal), then RTR= R-1R = 1.]
As with Example 3, we can reformulate the current example in a different manner. Suppose we define
C ≡ AB and specify that both A and B are contravariant rank-2 tensors. In this case, is C a contravariant
rank-2 tensor? If it were, we would have to have (A'B') = R(AB)RT from (5.7.1). But we showed above
that, since RTR ≠ 1, we end up with (A'B') ≠ R(AB)RT . Therefore, C ≡ AB is not a contravariant rank-2
tensor.
Could C be a covariant rank-2 tensor? If it were, we would need to have (A'B') = ST(AB)S from
(5.7.1). But above we show that A'B' = RA(RTR)BRT and this is completely different from (A'B') =
ST(AB)S. Thus, C is not a covariant rank-2 tensor.
Since C has two indices, the only way it could be a tensorial tensor is if it is either a contravariant or a
covariant rank-2 tensor, but we have just ruled out both these possibilities.
Therefore C ≡ AB is not a tensorial tensor of any kind whatsoever, even though A and B are tensorial
tensors.

106
 Chapter 7: Standard Notation

Matrix Rule #1. In developmental notation, if A and B are contravariant rank-2 tensors, the matrix
product AB is (in general) not a rank-2 tensor and is in fact not any kind of tensor. The equation C = AB
is not covariant. Mimicking the above discussion, the reader can show that the same conclusion applies to
C = ĀB, C = AB̄ and C = ĀB̄ : in none of these cases is C a tensor of any kind, and all these equations
are non-covariant. Similarly, the Rule applies to X = ABC or X = ABCD and so on. (7.8.5)

For this reason, we shall never ask how to transform an equation like X = ABC... from developmental to
standard notation. Equations which are non-covariant are simply of no interest, and can never describe a
physical relationship, as explained below in Section 7.15.

The attentive reader might ask: What about the equation g' = R g RT which has the form X = ABC. And
if g = 1, what about g' = RRT whose form is X = AB? In both these cases, the left hand side is a
contravariant rank-2 tensor. These equations do not violate the Matrix Rule #1 above because the matrices
R and RT are not tensors of any kind, as noted below (7.5.2). Furthermore, one does not ask whether g' =
R g RT is covariant or not because it is an equation relating objects in different spaces and not all objects
in the equation are tensors.

We now consider the notion of matrix multiplication using mixed rank-2 tensors. Since we never
introduced such mixed tensors in our developmental notation, we have this discussion entirely in the
Standard Notation. Consider

Cij = AikBkj . // implied sum on k (7.8.6)

The indices k have the right adjacency so one could think of this as being a matrix equation C = AB
where all three objects are "down-tilt rank-2 mixed tensors". Down-tilt just means the two indices are
tilting down like ij. We can ask again our questions of Example 4. If A and B are rank-2 tensors, is C =
AB covariant? And if we define C ≡ AB, is C a rank-2 tensor?
The answer to both questions is yes.
To show that AB = C is covariant, we start with the second line of (7.5.8) with Rbb' = Sb'b from
(7.5.13) :

M'ab = Raa' Sb'b Ma'b' (7.5.8)

and we apply this transformation rule to both A and B to get

A'ikB'kj = (Ria' Sb'k Aa'b') (Rka" Sb"j Ba"b")

= Ria' Sb'k Rka" Sb"j Aa'b'Ba"b" = Ria' (Sb'k Rka") Sb"j Aa'b'Ba"b"

= Ria' (SR)b'a" Sb"j Aa'b'Ba"b" = Ria' δ b'a" Sb"j Aa'b'Ba"b"

= Ria' Sb"j Aa'b'Bb'b" = Ria' Sb"j (AB)a'b" = Ria' Sb"j Ca'b"

= C'ij . // using (7.5.8) a third time with M = C (7.8.7)

107
 Chapter 7: Standard Notation

Thus we have shown that AB = C ⇒ A'B' = C' so our down-tilt matrix equation is covariant.
If we define C ≡ AB where A and B are down-tilt mixed rank-2 tensors, then C will be a rank-2
down-tilt tensor providing we can show that (A'B')'ab = Raa' Sb'b (AB)a'b = Raa' Sb'b Ca'b . But this
is just what was shown above (albeit with different indices), so yes, C is also a down-tilt mixed rank-2
tensor.
One way to clarify the intention of C = AB is to write the matrix equation as Cdt = AdtBdt where the
notation Adt means the down-tilt mixed rank-2 tensor having components Aij. In other words,

Cij = AikBkj ⇔ (Cdt)ij = (Adt)ik(Bdt)kj (7.8.8)

where the matrices Cdt, Adt and Bdt can be regarded as ordinary linear algebra matrices. For example,
the identity det(XY) = det(X)det(Y) then says

det(Cdt) = det(Adt)det(Bdt) or det(C**) = det(A**)det(B**) (7.8.9)

where we use wildcards to indicate the position of the indices of the elements which appear in the matrix
whose determinant we are computing.

It is easy to show that the conclusions reached above apply similarly to an all up-tilt matrix equation

Cij = Aik Bkj . // implied sum on k (7.8.10)

We thus arrive at:

Matrix Rule #2. In Standard Notation, it is reasonable to use matrix notation in the following two
situations involving mixed rank-2 tensors:

Cij = AikBkj Cdt = AdtBdt dt = down-tilt

Cij = AikBkj Cut = AutBut ut = up-tilt (7.8.11)

In special relativity the down-tilt matrix form is most often used, and one just writes C = AB without
bothering with the dt clarifying subscripts. This is consistent with the usual statement x'μ = Λμνxν to
describe a Lorentz transformation acting on the contravariant vector xν (where Λ = our R ).

Note: In the matrix equation Cij = AikBkj containing "down-tilt" mixed rank-2 tensors, the k index sum
"tilts up".

7.9 Matrix Inverse, Transpose and Determinant

Matrix Inverses

Consider the standard notation matrix equation AB=1 where (assuming det(A) ≠ 1) we can write B = A-1.
As demonstrated above, AB = 1 can only be a covariant equation if A and B are both down-tilt or both
up-tilt mixed rank-2 tensors. Then we are talking about either AdtBdt = 1dt or AutBut = 1ut, and the
corresponding Bdt = (A-1)dt and But = (A-1)ut, all these being matrix equations . Thus

108
 Chapter 7: Standard Notation

AdtBdt = 1dt Bdt = (A-1)dt AikBkj = δij Bij = (A-1)ij

AutBut = 1ut But = (A-1)ut AikBkj = δij Bij = (A-1)ij . (7.9.1)

In this context, we have shown that if A is a mixed rank-2 tensor, then (A-1) is a mixed rank-2 tensor as
well, assuming it exists.
In (7.5.11) we considered S = R-1 and R = S-1 and reached the conclusions shown in (7.9.1) for the
cases A = S and R = B. It happens that in this special case, S and R are not tensors, but the results are still
valid.
It was shown in (7.4.19) that 1dt is really gdt, the down-tilt form of the metric tensor g, and similarly
for 1ut :

1dt = gdt (1)ij = gij = δij = δi,j

1ut = gut (1)ij = gij = δij = δi,j . (7.9.2)

Thus first equation in (7.9.1) can be written in covariant form AdtBdt = gdt where all three matrices are
down-tilt mixed rank-2 tensors. And this is also true for AutBut = gut .

Transpose Matrices

It is possible to define the "covariant transpose" MT (italic T) of a matrix M in Standard Notation, and
things work out as follows, where M is any rank-2 tensor,

(MT)ab = Mba (RT)ab = Rba (ST)ab = Sba

(MT)ab = Mba (RT)ab = Rba (ST)ab = Sba
(MT)ab = Mba (RT)ab = Rba (ST)ab = Sba
(MT)ab = Mba (RT)ab = Rba (ST)ab = Sba . // covariant transpose (7.9.3)

The general rule is that, in each equation, the indices are reflected in a vertical line separating them.

The notation is "covariant" in that one can raise and lower indices at will in each equation and thereby
create a new valid equation. The drawback of this notation is that, when the indices are tilted, MT in
general differs from what we might call the "matrix transpose" MT of a matrix M. When one transposes a
matrix, one normally means to swap the rows and columns, and that means to swap the indices. One
would then write

(MT)ab = Mba (RT)ab = Rba (ST)ab = Sba

(MT)ab = Mba (RT)ab = Rba (ST)ab = Sba
(MT)ab = Mba (RT)ab = Rba (ST)ab = Sba
(MT)ab = Mba (RT)ab = Rba (ST)ab = Sba . // matrix transpose (7.9.4)

109
 Chapter 7: Standard Notation

The middle two lines (tilted indices) are very non-covariant in form since for example (MT)ab = Mba has
the up and down sense of index a and b not even matching. The use of the MT matrix transpose is in
theorems like det(M) = det(MT) where rows and columns are swapped without changing the determinant.

Comments:

1. The issue here is that one might have

(MT)ab = Mba ≠ (MT)ab = Mba .

If M is an x-space tensor and if the x-space metric tensor is g = 1, then up and down index positions don't
matter and our problem goes away,

(MT)ab = Mba = (MT)ab = Mba

so then MT = MT. (7.9.5)

2. For R and S: As shown in (7.5.9), g' acts on the first index of R while g acts on the second index.
Conversely as shown in (7.5.10), g acts on the first index of S while g' acts on the second index. So up
and down indices are the same for R and S only if both g = 1 and g' = 1. Only in this special case, which
implies a (local) rotation for x' = F(x), does one have RT = RT and ST = ST . This situation arises in
Appendix E where we have x" = FM(x) and R and S (there called M and N) are rotations. (7.9.6)

3. As it turns out, regardless of g and g', one has

det(M) = det(MT) = det(MT) (7.9.7)

for any index positions on the matrices. The traditional fact that det(M) = det(MT) is true for any kind of
square matrix, while the fact that det(M) = det(MT) is proven in (D.12.20). Since T and T are the same
when indices are both up or both down, the interesting case is when indices are tilted. Here is an outline,

det(MT) = det([MT]ij) = det(Mji) = det(gjj'Mj'i'gi'i) = det(gdnMdtgup)

= det(gdn)det(Mdt)det(gup) = g det(Mdt) (1/g) = det(Mdt) = det(Mij) = det(M) .

4. As shown in the next few paragraphs, in either up-tilt or down-tilt notations one can write,

RRT = RTR = 1 SST = STS = 1 RS = SR = 1

RT = R-1 = S ST = S-1 = R . (7.9.8)

Thus, both the R matrix and S matrix are "covariant real-orthogonal" for any underlying transformations
x' = F(x), even when R and S are not rotation matrices. Since the matrix elements are real, R and S are
also "covariant unitary".

110
 Chapter 7: Standard Notation

Restatement of Orthgonality Conditions

We shall generally avoid using the covariant transpose notation since it leads to results which can be
confusing although correct, but there will be times when we use it. To restate the potential confusion, in
developmental notation the statement RRT = 1 (or RT = R-1, meaning "real orthogonal") implies that R is
a "rotation", where we include in this term the possibility of axis reflections. But in Standard Notation,
RRT = 1 is valid for any matrix R associated with a general transformation x' = F(x), rotations and non-
rotations alike. To see why this is so, just write out orthogonality rules from (7.6.4) :

#4 Rab Rcb = δac ⇒ Rab (RT)bc = δac ⇒ RRT = 1 (up-tilt) .

#3 Rab Rcb = δac ⇒ Rab (RT)bc = δac ⇒ RRT = 1 (down-tilt) .

#2 Rba Rbc = δac ⇒ (RT)ab Rbc = δac ⇒ RTR = 1 (up-tilt) .

#1 Rba Rbc = δac ⇒ (RT)abRbc = δac ⇒ RTR = 1 (down-tilt) . (7.9.9)

The equations on the right are just restatements of the orthogonality rules. The same equations are valid
with R→ S, for example

RRT = 1 ⇒ (RRT)-1 = 1 ⇒ (RT)-1R-1 = 1 ⇒ (R-1)TR-1 = 1 ⇒ STS = 1 .

To totally distinguish the above situations, we could use this rather unpleasant notation,

[RRT = 1]DN ⇔ R is a "rotation" DN = Developmental Notation

[RRT = 1]SN,ut for any R associated with F SN,ut = Standard Notation, up-tilt

[RRT = 1]SN,dt for any R associated with F SN,dt = Standard Notation, down-tilt (7.9.10)

But such notations are not necessary if one understands that an equation like RRT = 1 implies all up-tilt or
all down-tilt matrices.

How to recognize a rotation

How then does one recognize a rotation matrix directly in Standard Notation? We translate using the rule
Rik→ Rik shown in (7.5.2),

[RRT= 1]DN ⇒ RikRjk = δi,j → RikRjk = δi,j .

A simpler alternate form for recognizing a rotation can be obtained as follows,

111
 Chapter 7: Standard Notation

RikRjk= δi,j // now apply Sai to both sides

SaiRikRjk= Saiδi,j

(SR)akRjk = Saj // but SR = 1

δakRjk = Saj = Rja // right side from (7.5.13)

Rja = Rja // if R is a rotation

Therefore, in Standard Notation a "rotation" can be identified in either of these two ways,

RikRjk = δi,j or Rja = Rja // "rotation" . (7.9.11)

Matrix form of the transformation rule for a rank-2 tensor

Instead of trying to use this covariant transpose T superscript in the standard notation, we can get rid of T
in the developmental notation and then do our translation to standard notation. For example we start in
developmental notation,

M' = R M RT ⇒ M'ad = RabMbc(RT)cd = RabMbcRdc = RabRdcMbc (7.9.12)

Then we make the conversion using (7.5.2) Rik → Rik and (7.4.1) Mab → Mab :

M'ad = RabRdcMbc → (M')ad = RabRdcMbc (7.9.13)

and this then is the Standard Notation rule for the way a contravariant rank-2 tensor transforms, as was
shown in the first line of (7.5.8).

However, notice that, where we make use of the "tilt reversal theorem" (7.4.15),

(M')ad = RabRdcMbc // covariant, so can lower index d on both sides

so
(M')ad = RabRdcMbc = RabRdcMbc = RabMbcRdc = RabMbc(RT)cd = (RMRT)ad . (7.9.14)

Thus we can write M' = RMRT in standard notation provided we assume that we are using all down-tilt
matrices. The same equation results if all matrices are up-tilt:

(M')ad = RabRdcMbc
so
(M')ad = RabRdcMbc = RabMbcRdc = RabMbc(RT)cd = (RMRT)ab (7.9.15)

We then arrive at this conclusion:

112
 Chapter 7: Standard Notation

Fact: The transformation rule for a rank-2 tensor M can be written in matrix notation in the Standard
Notation as

M' = RMRT all up-tilt or all down-tilt (7.9.16)

provided it is understood that all matrices are up-tilt or all are down-tilt. The result is very similar to what
it was in Developmental Notation: M' = RMRT. The use of the covariant transpose thus causes this
equation to maintain its form when we go to Standard Notation.

Transpose in Dirac Notation

In the Dirac notation dot products are represented as a • b = <a | b> = b • a = <b | a>. In general Dirac
theory (as used in quantum mechanics) one has <a|b> = <b|a>*, but all our scalar products are real so we
can ignore complex conjugation. Consider then,

<a | M | b> ≡ <a | M b> = a • (Mb) = ai(Mb)i = ai[ Mijbj] = ai Mij bj

= bj Mij ai = bj (MT)ji ai = bj [MTa]j = b • (MTa) = <b| MTa> = <b| MT |a> . (7.9.17)

This proves:

Fact: In Dirac notation <a | M | b> = <b | MT | a> for any rank-2 tensor M, where T is the covariant
transpose. (7.9.18)

If a and b are vectors in Cartesian space with g = 1, then <a | M | b> = <b| MT |a> since then MT = MT.

A fuller description of Dirac notation is presented in (E.7.4).

Summary of Covariant Transpose Equations

We summarize the above facts which relate to the covariant transpose in Standard Notation,

det(M) = det(MT) determinant matrix transpose theorem (7.9.7)

RS = SR = 1 so R-1 = S and S-1 = R relation between S and R from their definition (7.5.11)
RRT = RTR = 1 ⇒ RT = R-1 = S orthogonality rules (R is covariant real orthog.) (7.9.8)
SST = STS = 11 ⇒ ST = S-1 = R orthogonality rules (S is covariant real orthog.) (7.9.8)
M' = RMRT how a (tilted) rank-2 tensor transforms (7.9.16)
<a | M | b> = <b | MT | a> Dirac matrix elements of operator M (7.9.17) (7.9.19)

In these equations MT is the covariant transpose shown in (7.9.3). It is assumed that all matrix products
involve matrices which are either all down-tilt or all up-tilt. The various orthogonality rules like RRT = 1
are valid for any R matrix, not just for a rotation.

113
 Chapter 7: Standard Notation

Determinant of a Matrix

In developmental notation one writes

det(A) = εabc... A1aA2bA3c..... = εabc... Aa1 Ab2Ac3..... (7.9.20)

where the Aij are components of the contravariant rank-2 tensor A and where ε is the permutation tensor
discussed above in Section 7.7.
We have argued above that the notion of a rank-2 tensor being a matrix in Standard Notation is only
viable for mixed rank-2 tensor of either the down-tilt or up-tilt variety. Thus, the matrix determinants of
interest in Standard Notation would be these:

det(Adt) = det(A**) = εabc... A1a A2bA3c..... = εabc... Aa1 Ab2Ac3..... (7.9.21)

det(Aut) = det(A**) = εabc... A1a A2bA3c..... = εabc... Aa1 Ab2Ac3..... (7.9.22)

Determinants of a rank-2 tensor A rarely appear in this document, but they do appear for the non-tensor
objects R and S as shown in (7.5.17), as needed for the Jacobian in (5.12.6).
For more information on determinants of rank-2 tensors see Section D.12.

7.10 Tensors of Rank n, direct products, Lie groups, symmetry and Ricci-Levi-Civita

The most general tensor of rank n (aka order n) will have some number s of contravariant indices and
then some number n-s of covariant indices. If s = n, the tensor is pure contravariant, and if s = 0, it is pure
covariant, otherwise it is "mixed" (as opposed to "pure"). The transformation of the tensor under F will
show a factor Raa' for each contravariant index, and a factor Sa'a ( = Raa' by (7.5.13) ) for each
covariant index, as illustrated by this example :

T 'abcde = Raa' Rbb' Rcc' Sd'd Se'e Ta'b'c'd'e'

T 'abcde = Raa' Rbb' Rcc' Rdd' Ree' Ta'b'c'd'e' . (7.10.1)

Note that for Raa' and Raa' the second index is the summation index, but for Sa'a it is the first index.

A rank-n tensor always transforms the way an outer product of n vectors transforms if those vectors have
indices which type-match those of the tensor. In the above case, an object that would transform the same
as Tabcde would be

AaBbCcDdEe . (7.10.2)

A tensor of rank-n has 2n tensor objects in its family since each index can be up or down. For example,
the tensor T above is one of 25 = 32 tensors one can form. Of these, one is pure covariant and one is pure
contravariant and 30 are mixed.
If any of these tensors is a tensor field, such as Tabcde(x), then of course all family members are
tensor fields.

114
 Chapter 7: Standard Notation

Direct Products. Consider again the outer product of vectors AaBbCcDdEe. The transformation A'a =
Raa'Aa' occurs in an N-dimensional contravariant vector space we shall call R . In this space one could
establish a set of basis vectors, and of course there are rules for adding vectors and so on. Transformation
B'a = RacBc occurs in an identical copy of the space R, but transformation D'd = Sd'dDd' = Rdd'Dd'
occurs in a covariant version of R we call R̄ . Since dot products (inner products) have been established
for vectors in these spaces, they can be regarded as full blown Hilbert Spaces with the caveats of Section
5.10.
The transformation of the outer product object, as already noted, is given by

A'aB'bC'cD'dE'e = Raa' Rbb' Rcc' Rdd' Ree' Aa'Bb'Cc'Dd'Ee' (7.10.3)

and one can consider the operator Raa' Rbb' Rcc' Rdd' Ree' as a transformation element in a so-called
direct product space which in this case would be written

Rdp = R ⊗ R ⊗ R ⊗ R̄ ⊗ R̄ (7.10.4)

One could then define

(Rdp)abcde ; a'b'c'd'e' ≡ Raa' Rbb' Rcc' Rdd' Ree' (7.10.5)

so that

A'aB'bC'cD'dE'e = (Rdp)abcde ; a'b'c'd'e' Aa'Bb'Cc'Dd'Ee' (7.10.6)

and of course this would apply to any tensor of the same index configuration, such as

T 'abcde = Rdpabcde ; a'b'c'd'e' Ta'b'c'd'e' . (7.10.7)

This suggests a definition of "tensor" as follows" : tensors are those objects that are transformed by all
possible direct product representations formable from the two fundamental vector representations R and
R̄. To this set of spaces one would add the identity space 1 to handle tensorial scalars.
Appendix E continues this direct product discussion in terms of the basis vectors that form a complete
set for a direct product space such as Rdp and shows how to expand tensors on such bases.

Lie Groups. The direct product notion is just a formalism, but the formalism has some implications when
the space R is associated with a "representation" of a Lie group. In this case, a direct product Rdp = R ⊗
R can be written as a sum of "irreducible" representations of that group. What this means is that the
transformation elements of Rdp and the objects Tab can be shuffled around with linear combinations so
that (Rdp)aba'b', when thought of as a matrix with columns labeled by N2 ab possibilities and rows
labeled by the N2 a'b' possibilities, appears in "block diagonal form" with all zeros outside the blocks. In
this case, the shuffled components of tensor Tab can be regarded as a non-interacting assembly of pieces
each of which transforms according to one of those blocks of the shuffled (Rdp)aba'b'.
The most famous example occurs with N=3 and the rotation group SU(2) in which case R(1) ≡ R can
be decomposed according to R(1) ⊗ R(1) = R(2) ⊕ R(1) ⊕ R(0) where the ⊕ symbols indicate this

115
 Chapter 7: Standard Notation

block diagonal form. In this case the blocks are 5x5, 3x3 and 1x1, fitting onto the diagonal of the 9x9
matrix area. The numbers L = 0,1,2 here label the rotation group representations and that label is
associated with angular momentum. The elements of the 5x5 block are called D(2)M,M'(φ,θ,ψ) where
M,M' = 2,1,0,-1.-2, and where φ,θ,ψ are the "Euler angles" which serve to label a particular rotation. This
D(2)object is the L=2 matrix representation of the rotation group. Taking two vectors A and B, one can
identify A•B as the combination transforming according to R(0) ("scalar") and AxB ( linearly combined)
as that transforming as R(1) ("vector"). The traceless symmetric matrix AiBj - δi,jA•B has 5
independent elements associated with R(2) ( "quadrupole").
This whole reduction idea can be applied to larger direct products such as R(1) ⊗ R(1) ⊗ R(1) and
tensor components Tabc.
The Standard Model of elementary particle physics is chock full of direct products of this nature,
where the idea of rotational symmetry is extended to other kinds of "internal" symmetry, spin and isospin
being two examples. Representations of the Lie symmetry group SU(3) are associated with quarks which
are among of the fundamental building blocks of the Standard Model.
The group discussion above can be applied generally to quantum physics. The basic idea is that if "the
physics" (the Hamiltonian or Lagrangian) describing some quantum object is invariant under a certain
symmetry group (such as rotational symmetry or perhaps some discrete crystal symmetry), then the
quantum states of that object can be classified according to the representations of that group. The Bohr
hydrogen atom "physics" H ~ ∇2-1/|r| has perfect rotation group symmetry and is also symmetric about
the axis (angle ψ) from center to electron (no spin). The representation functions then must have M' = 0,
and then D(L)M,0(φ,θ,ψ) ~ YLM (θ,φ), the famous spherical harmonics that describe the "orbitals" which
have mystified first-year chemistry students for the last 100 years.

Historical Note: Ricci and Levi-Civita (see Refs) referred to rank-n tensors as "systems of order n" and
did not include mixed tensors in their 1900 paper. Nor did they use the Einstein summation convention,
since Einstein thought of that later on. They did use the up and down index notation pretty much as it is
used today, though the up indices are enclosed in parenthesis. Here is a direct quote from the paper where
the nature of the contravariant and covariant tensors is described (with crude translation below for non-
French readers). Equation (6) had typos which some thoughtful reader corrected: the y subscripts should
be r's and the x subscripts should be s's. In our notation ∂xs/∂yr → ∂xs/∂x'r = Ssr = Rrs.

116
 Chapter 7: Standard Notation

(7.10.8)
We will say that a system of order m is covariant (and in this case we will designate its elements by the
symbol Xr1,r2....) (r1,r2.... can each take all the values 1...n), if the elements Yr1,r2.... of the
transformed system are given by the formulas (6) .

We will designate on the contrary by the symbols X(r1,r2....) the elements of a contravariant system,
which is to say of a system where the transformation is represented by the formulas (7) ,

the elements X and Y being related respectively to (presumably "are functions of") the variables x and y.

Their y is our x', and their n is our N. Notice that the indices on the coordinates themselves are taken
down, contrary to current usage. They do not explain why the words contravariant and covariant are used.
In our notation, the two equations above would be written

M'abc... = Raa'Rbb'....... Ma'b'c'... = Σprimed Ma'b'c'... Raa'Rbb'....... (6)

M'abc... = Raa' Rbb'....... Ma'b'c'... = Σprimed Ma'b'c'... Raa' Rbb''....... (7) (7.10.9)

7.11 The Contraction Tilt-Reversal Rule

In some complicated combination of multiple tensors, imagine there is somewhere a pair of summed
indices where one is up and the other is down. As noted above (7.4.15), such a sum is called a
contraction. The contracted indices could be on the same object or they could be on different objects. We
depict this situation with the following symbolic notation,

[-----a---------a----] (7.11.1)

where the dashes indicate indices that we don't care about and which won't change -- each one could be
up or down. We know we can reverse the tilt this way, as in (7.4.11),

117
 Chapter 7: Standard Notation

[-----a---------a----] = gab gac [-----b---------c----] (7.11.2)

where the first g raises the index b to a, and the second g lowers the index c to a. But the two g's are
inverses as in (7.4.1), gab gac = gba gac = δb,c, which at once gives the desired result (index b→a)

[-----a---------a----] = [-----a---------a----] . // the Contraction Tilt-Reversal Rule (7.11.3)

A notable example of course is this:

AaYa = AaYa = " A•Y" // or perhaps " A.Y " as noted in Section 5.10 (7.11.4)

When is index tilt-reversal allowed and when is it not allowed?

It is always allowed when both indices of the tilted contraction are valid tensor indices. Consider these
four examples to be discussed below:

RabAb = RabAb Proof: RabAb = Racgcb gbdAd = gcbgbdRacAd = δcdRacAd = RacAc

AaRab ≠ AaRab Proof: AaRab = gacAc Rdb g'da = gac g'ad Ac Rdb ≠ AaRab

Aa∂a = Aa∂a Proof: Aa∂af = gacAc gad∂df = gacgadAc∂df = δcdAc∂df = Ac∂cf (7.11.5)

∂aAa ≠ ∂aAa Proof: ∂aAa = (gac∂c)(gadAd) = gacgad(∂cAd) + gac(∂cgad)Ad

= δcd(∂cAd) + gac(∂cgad)Ad = ∂cAc + (∂agad)Ad ≠ ∂cAc

In the first example, since Rab is not a tensor, one is on dangerous ground doing the tilt reversal, but it
happens to work because the second index is associated with metric tensor gij which is the same metric
tensor that raises and lowers indices of Ab. In the second example, the tilt-reversal fails because the first
index of Rab is associated with the x'-space metric tensor g'ij. In the third example, both indices are valid
tensor indices (with the same metric tensor).
The fourth example shows a failure of the tilt-reversal rule and this example is very important. The
inequality becomes an equality only if the underlying transformation F is linear so that R and S and g are
then constants independent of position. For general F, such as the F involved in curvilinear coordinate
transformations, the object ∂aAb is not a rank-2 tensor and so the object ∂aAa does not represent
contraction of two true tensor indices and therefore the "contraction tilt-reversal rule" does not apply. The
neutralization rule of Section 7.12 below also does not apply for this same reason, so ∂aAa does not
transform as a scalar under general F.

Here is one more example along the lines of the fourth example above that shows the potential danger of
reversing a tilt when it is not justified. In this example, we use notation to be introduced in Section 7.16
below. Consider the equation,

Va = εabcBb;c (1) // valid

where

118
 Chapter 7: Standard Notation

Bb = a tensorial vector
Bb;c = ∂cBb – ΓnbcBn = the covariant derivative of vector Bb = a rank-2 tensor
εabc = a rank-3 tensor density (weight -1) (the Levi-Civita tensor)
Va = a vector density (weight -1) (7.11.6)

With regard to Bb;c : (a) in comma notation one writes ∂cBb = Bb,c ; (b) Γnbc = Γncb .

Since all indices on equation (1) are tensor indices, one can lower index a and reverse the b and c tilts to
get
Va = εabcBb;c . (2) // valid (7.11.7)

Now go back to equation (1). Because εabc is antisymmetric on b and c, whereas Γnbc is symmetric on b
and c, one can write εabcBb;c = εabcBb,c since the Γ term vanishes by symmetry. Thus one gets

Va = εabcBb,c . (3) // valid (7.11.8)

Were one to blindly lower a and reverse the b and c tilts, one would get

Va = εabcBb,c . (4) // NOT valid (7.11.9)

The reason for "not valid" is that the reversal of the b tilt is not justified. That is,

Va = εabcBb,c = εabc∂cBb = gbb' εab'c ∂c(gbb"Bb")

⇒ Va = gbb' εab'c ∂c(gbb"Bb") = gbb' εab'c ∂c(gbb"Bb") // c tilt reversal is OK

= gbb'gbb" εab'c(∂c Bb") + gbb' εab'c (∂c gbb") Bb"

= δb'b" εab'c(∂c Bb") + gbb' εab'c (∂c gbb") Bb"

= εabc(∂cBb) + gbb' εab'c (∂c gbb") Bb"

= εabcBb,c + gbb' εab'c (∂c gbb") Bb" = εabcBb,c + extra term! (7.11.10)

It is due to this extra term that (4) is not valid. Basically this is the same as the fourth example above, but
the situation is embedded in a more complicated environment (extra tensors, tensor densities, comma
notation, covariant derivatives, other tilted indices, etc). One way to summarize the example is this:

(Va = εabcBb,c) ⇔ (Va = εabcBb;c) ⇔ (Va = εabcBb;c) ⇔

/ (Va = εabcBb,c) (7.11.11)

7.12 The Contraction Neutralization Rule

A contracted index pair plays no role in how a tensor object transforms; the two contracted indices
neutralize each other, as we now show. (7.12.1)

119
 Chapter 7: Standard Notation

First, recall that the indices on a general rank-n tensor (perhaps formed from several tensors) transform
the same way that an outer product of n vectors transforms, where the vector index types match those of
the tensor. The vectors transform this way,

V'a = RabVb V'a = SbaVb . (7.5.3) and (7.5.4)

We take the same "big object" [-----a---------a----] appearing in (7.11.1) and now ask how it transforms.

Below the X's represent either R or S factors for the dash indices (each of which might be up or down):

[-----a---------a----]' = XXXXX Rab XXXXXXXXX Sca XXXX [-----b---------c----]

= Sca Rab XXXXX XXXXXXXXX XXXX [-----b---------c----]

= δcb XXXXX XXXXXXXXX XXXX [-----b---------c----] // RS = 1

= XXXXX XXXXXXXXX XXXX [-----a---------a----] (7.12.2)

where now the only X's left are for the other indices. Again we look at our canonical example,

AaYa = AaYa = A•Y . (7.4.14)

The contracted vector indices neutralize each other and the resulting object transforms as a scalar.

Here are some examples of tensor transformations with 0,1 and 2 index pairs contracted:

T 'abcde = Raa' Rbb' Rcc' Sd'd Se'e Ta'b'c'd'e' // no pairs contracted

T 'abcae = Rbb' Rcc' Se'e Ta'b'c'a'e' // index a contracted
T 'abcab = Rcc' Ta'b'c'a'e' // index a and index b contracted
Q' = Q where Q = AaYa and Q' = A'aY'a // index a contracted (7.12.3)

This shows the idea that one can take a larger tensor like Tabcde and form from it smaller (lower rank)
tensors by contracting tilted pairs of indices. In the above example list we really have

Dbce ≡ Tabcae = a mixed rank-3 tensor

Ec = Tabcab = a contravariant vector (rank-1 tensor)

Q = AaYa = a scalar (rank-0 tensor) . (7.12.4)

It is similarly possible to build larger tensors from smaller ones, for example

Zabcde = Va We gab Lc (7.12.5)

which goes under the same rubric "outer product" mentioned earlier.

120
 Chapter 7: Standard Notation

7.13 The tangent and reciprocal base vectors and expansions on same

Tangent and reciprocal base vectors

Here are some basic translations. In the first four lines, n is a label, i is a tensor index:

(en)i → (en)i // contravariant index i

( ē n)i → (en)i // covariant index i

(En)i → (en)i // contravariant index i

(Ēn)i → (en)i // covariant index i

(en)i = Sin (3.2.5) → (en)i = Sin = Rni // contravariant index i

(Ēn)i = Rni (6.3.2) → (en)i = Sin = Rni // covariant index i

(En)i = Rnkgki (6.1.4) → (en)i = Rnkgki = Rni // contravariant index i

Σn(Ēn)a(en)b = δb,a (6.2.16) → Σn(en)a(en)b = δb,a // completeness

Σn Ēn enT = 1 (6.2.24) → Σn en enT = 1 // completeness (matrix form) (7.13.1)

As noted earlier, writing a vector in bold such as en is not enough to say whether the vector is
contravariant or covariant. Here we use an index position marker * on the right. For example, (e1)* is a
column vector of contravariant components, matching Sin = (en)i from (7.13.1).

S = [e1, e2, e3 .... eN ] (3.2.7) → S** = [(e1)*, (e2)*, (e3)* .... (eN)*] (7.13.2)

R = [Ē1, Ē2, Ē3 .... ĒN ]T (6.3.3) → R** = [(e1)*, (e2)*, (e3)* .... (eN)*]T (7.13.3)

The relationship between en and en is very simple. Start with (6.1.2) and use (7.4.1) g'ab → g'ab to get the
first result below. Then apply g'mn to the first result so g'mnen = g'mn g'ni ei = δmi ei = en which is the
second result:

En ≡ g'ni ei → en = g'ni ei and en = g'ni ei . (7.13.4)

For either contravariant or covariant indices (indices are not shown!), g'ni raises the label on ei, and
inverting one finds that g'ni lowers the label on ei. This fact makes things easy to remember. Using the
fact (3.2.6) that ei = ∂'ix , one has g'ni ei = g'ni ∂'ix = ∂'nx so the above line can be expressed as

En ≡ g'ni ei → en = ∂'nx and en = ∂'nx . (7.13.5)

121
 Chapter 7: Standard Notation

The dot products (6.2.4) translate this way:

en • em = ḡ'nm → en • em = g'nm = ∂'nx • ∂'mx |en| = g'nn = h'n

En • em = δn,m → en • em = δnm = ∂'nx • ∂'mx

En • Em = g'nm → en • em = g'nm = ∂'nx • ∂'mx |en| = g'nn (7.13.6)

The "labels" on the base vectors behave in this dot product structure the same way that up and down
"indices" behave. This is the motivation for En → en . Thus, the three final equations can be regarded as
the same equation en • em = g'nm where we can raise either or both indices/labels to get the other
equations. For example, en • em = g'nm = δnm .
The dot product (6.1.7) becomes

En • um = Rnm → en • um = Rnm // = <en | um> in bra-ket notation (App E (g)) (7.13.7)

The matrix Rnm is a "basis change matrix" between basis uk and basis ek .

Inverse tangent and reciprocal base vectors

Using the rules given above in (6.5.2),

g'↔ g R↔S en → u'n e'n → un En → U'n E'n → Un (6.5.2)

we can obtain from (7.13.1) and (7.13.6) the corresponding results for the inverse tangent and reciprocal
base vectors:

122
 Chapter 7: Standard Notation

(u'n)i → (u'n)i // contravariant index i

(ū'n)i → (u'n)i // covariant index i

(U'n)i → (u'n)i // contravariant index i

(Ū'n)i → (u'n)i // covariant index i

(u'n)i = Rin → (u'n)i = Rin = Sni // contravariant index i

(Ū'n)i = Sni → (u'n)i = Rin = Sni // covariant index i

(U'n)i = Snkg'ki → (u'n)i = Snkg'ki = Sni // contravariant index i

¯ n)a(u'n)b = δb,a (6.2.16) →

Σn(U' Σn(u'n)a(u'n)b = δb,a // completeness

R = [u'1, u'2, u'3 .... u'N ] → R** = [(u'1)*, (u'2)*, (u'3)* .... (u'N)*]
S = [Ū'1, Ū'2, Ū'3 .... Ū'N ]T → S** = [(Ū'1)*, (Ū'2)*, (Ū'3)* .... (Ū'N)*]T

U'n ≡ gni u'n → u'n = gni u'i and u'n = gni u'i

u'n • u'm = ḡ nm → u'n • u'm = gnm

U'n • u'm = δn,m → u'n • u'm = δnm
U'n • U'm = gnm → u'n • u'm = gnm (7.13.8)

Summary table.

The summary table given in (6.5.9) was this

x'-space x-space (6.5.9)

axis-aligned basis vectors e'n un (e'n)i= δn,i (un)i= δn,i
dual partners to the above E'n Un (E'n)i = g'ni (Un)i = gni
tangent base vectors u'n en (u'n)i= Rin (en)i = Sin
reciprocal base vectors U'n En (U'n)i = g'ia Sna (En)i = gia Rna
= gnaRia = g'naSia
We translate this entire table into Standard Notation:

x'-space x-space (7.13.9)

axis-aligned basis vectors e'n un (e'n)i= δni (un)i = δni
dual partners to the above e'n un (e'n)i = g'ni (un)i = gni
tangent base vectors u'n en (u'n)i = Rin (en)i = Sin= Rni
reciprocal base vectors u'n en (u'n)i = g'ia Sna (en)i = gia Rna
(u'n)i = Sni (en)i = Rni

123
 Chapter 7: Standard Notation

x-space expansions

The x-space expansions of (6.6.9)

V = V1 u1 + V2 u2 +... = ΣnVn un where Un • V = Vn Un = gni ui

V = V̄1 U1 + V̄2 U2 +... = ΣnV̄n Un where un • V = V̄n
V = V'1 e1 + V'2 e2 +... = Σn V'n en where En • V = V'n En = g'ni ei
V = V̄'1 E1 + V̄'2 E2 +... = Σn V̄'n En where en • V = V̄'n (6.6.9)

translate into the following :

V = V1 u1 + V2 u2 +... = ΣnVn un where un • V = Vn un = gni ui

V = V1 u1 + V2 u2 +... = ΣnVn un where un • V = Vn
V = V'1e1 + V'2e2 +... = Σn V'n en where en • V = V'n en = g'ni ei
V = V'1e1 + V'2 e2 +... = Σn V'n en where en • V = V'n (7.13.10)

x'-space expansions

Similarly, the x'-space expansions of (6.6.15)

V' = V'1 e'1 + V'2 e'2 +... = ΣnV'n e'n where E'n • V' = V'n E'n = g'ni e'i
V' = V̄'1 E'1 + V̄'2 E'2 +... = ΣnV̄'n E'n where e'n • V' = V̄'n
V' = V1u'1 + V2u'2 +... = Σn Vn u'n where U'n • V' = Vn U'n = gni u'i
V' = V̄1U'1 + V̄2U'2 +... = Σn V̄n U'n where u'n • V' = V̄n (6.6.15)

translate into the following :

V' = V'1 e'1 + V'2 e'2 +... = ΣnV'n e'n where e'n • V' = V'n e'n = g'ni e'i
V' = V'1 e'1 + V'2 e'2 +... = ΣnV'n e'n where e'n • V' = V'n
V' = V1u'1 + V2u'2 +... = Σn Vn u'n where u'n • V' = Vn u'n = gni u'i
V' = V1u'1 + V2u'2 +... = Σn Vn u'n where u'n • V' = Vn (7.13.11)

Summary of all expansions:

Using implied sum notation, we can now summarize the eight expansions above, plus the unit vector
expansion onto ^
e n, on just two lines :

V = Vn un = Vn un = V'n en = V'n en = V'n ^

en // x-space expansions, ⇒ V'n = h'nV'n
n n n n
V' = V' e'n = V'n e' = V u'n = Vn u' . // x'-space expansions (7.13.12)

In all cases one sees a tilted index summation where one index is a vector index and the other is a basis
vector label. Half the forms shown above can be obtained from the others by just "reversing the tilt". The
power of the Standard Notation makes itself felt in relations like these.

124
 Chapter 7: Standard Notation

Due to this tilt situation, sometimes a basis like {en} appearing in V = V'n en is called a "covariant
basis" while the basis {en} appearing in V = V'n en is called a "contravariant basis".
Corresponding expansions of higher rank tensors are presented in Section 7.17 below.
If V is a tensor density of weight W (see Appendix D and E) the rule for adjusting the above
expansions is to make the replacement V'n → JW V'n and V'n → JW V'n where J is the Jacobian of (5.12.6).

7.14 Comment on Covariant versus Contravariant

Consider this expansion for a vector V in x-space,

V = Vnbn Vn = V • bn (7.14.1)

where bn is some basis having dual basis bn where as usual bn • bm = δnm, see Section 6.2. Imagine
taking Vn → V'n = Rnm Vm and bi → bi' = Qij bj. What Q would cause the following to be true?

V = Vnbn = V'nb'n . (7.14.2)

In other words, how does one transform that basis bn such that the vector V remains unchanged if Vn is
transformed contravariantly? The answer to this question is that Qij = Rij since then (using R
orthogonality rule #2 of (7.6.4) )

V'nb'n = [Rnm Vm][ Rnj bj] = (Rnm Rnj) Vm bj = δmj Vm bj = Vj bj = Vnbn . (7.14.3)

Compare then the transformation of Vn with that of the basis bn:

V'n = Rnm Vm
bn' = Rnm bm (7.14.4)

The Vm vector components transform with Rnm but the basis vectors have to transform with Rnm to
maintain the invariance of the vector V. One varies with the down-tilt R, while the other varies with the
up-tilt R, so the two objects are varying against each other in this tilt sense. They are "contra-varying", so
one refers to the components Vm as contravariant components with respect to the basis bm .

If one starts over with Vn components and the bn "dual" (reciprocal) expansion vectors and asks for a
solution to this corresponding problem,

V = Vnbn = V'nb'n (7.14.5)

one finds not surprisingly that the dual basis must vary as bn' = Rnm bm and then one has

125
 Chapter 7: Standard Notation

V'n = Rnm Vm
bn' = Rnm bm (7.14.6)

which is the previous result with all indices up↔down. Comparing the tilts, one would say that the Vm
again "contra vary" with the way the bm vary to maintain invariance of V. But one does not care about the
dual basis, one cares about the basis, so relative to the basis bn one has

V'n = Rnm Vm
bn' = Rnm bm . (7.14.7)

If the basis bm is varied as shown in (7.14.7), then the dual basis bm varies as in (7.14.6) and V of (7.14.5)
remains invariant. Comparing now the way the Vm transform with the way the basis vectors bm transform
in (7.14.7), one sees that both equations have the same tilted Rnm. They are "co-varying", so one refers to
the components Vm as covariant components with respect to the basis bm .

7.15 The Significance of Tensor Analysis

"Why is tensor analysis important?", the reader might ask in the midst of this storm of index shuffling.
Now is a good time to answer the question. Consider the following sample equation in x-space, where the
fields Q, H, T and B may or may not be tensor fields:

Qadc(x) = Hab(x)Tbc(x) Bd(x) . (7.15.1)

Notice that when contracted indices are ignored, the remaining indices have the same type on both sides.
If the various objects really were tensors, one would say this was a "valid tensor equation" based on the
index structure just described.

One says that an equation is "covariant with respect to transformation x' = F(x)" if the equation has
exactly the same form in x'-space that it has in x-space , which for our example would be

Q'adc(x') = H'ab(x')T 'bc(x') B'd(x') . (7.15.2)

Here the word "covariant" has a new meaning, different from its being a type of vector or index. The
meaning is related in the sense that, comparing the above two equations, everything has "moved" in the
same manner ("co-varied") under the transformation. (Some authors think the word "invariant" is more
appropriate; Ricci and Levi-Civita used the term "absolute"; continuum mechanics uses the term "frame-
indifferent".)
If the objects Q, H, T and B are tensors under F, then covariance of any valid tensor equation like the
one shown above is guaranteed!!
The reason is that, once the contracted indices on the two sides are ignored according to the
"contraction neutralization rule" (7.12.1), the objects on the two sides of the equation have the same
indices which are of the same type, so both sides are tensors of the same type, and therefore both sides
transform from x-space to x'-space in the same way. If one starts, for example, with the primed equation
and installs the known transformations for all the pieces, one ends up with the unprimed equation.

126
 Chapter 7: Standard Notation

If this explanation is not convincing, a brute force demonstration can perhaps help out. The following
is also a good exercise is using the two tilt forms of the R matrix. Recall from (7.5.13) that Sba = Rab and
that SR = 1 is replaced by the various orthogonality rules (7.6.4).

We shall process the primed equation into the unprimed one :

Q'adc(x') = H'ab(x')T 'bc(x') B'd(x') // x'-space equation (7.15.3)

[Raa'Rdd'Rcc'Qa'd'c'(x)] = [Raa'Rbb' Ha'b'(x)] [Rbb"Rcc' Tb"c'(x) ] [Rdd'Bd'(x)]

= Raa' Rdd' Rcc'(Rbb' Rbb") Ha'b'(x) Tb"c'(x)Bd'(x) . (7.15.4)

Using orthogonality rule #1 of (7.6.4) we continue

= Raa' Rdd' Rcc'(δb'b") Ha'b'(x) Tb"c'(x) Bd'(x)

= Raa' Rdd' Rcc' Ha'b'(x) Tb'c'(x) Bd'(x) (7.15.5)

so that, using the fact that Q is a tensor to replace Q' on the left side of (7.15.3),

(Raa'Rdd'Rcc') Qa'd'c'(x) = (Raa' Rdd' Rcc') Ha'b'(x) Tb'c'(x) Bd'(x) . (7.15.6)

Now apply the Cancellation Rule (7.6.12) three times to conclude that

Qa'd'c'(x) = Ha'b'(x) Tb'c'(x) Bd'(x) (7.15.7)

and then remove all primes on indices to get

Qadc(x) = Hab(x)Tbc(x) Bd(x) . // x-space equation (7.15.8)

Thus it has been shown that, if all the objects transform as tensors, the equation is covariant.

Tensor density equations are also covariant. As discussed in Appendix D, a tensor density of weight W is
a generalization of a tensor which has the same transformation rule as a regular tensor, but there is an
extra factor of J-W on the right hand side of the rule, where J is the Jacobian J = detS. For example,

Q'adc(x') = J-WQ Raa'Rdd'Rcc'Qa'd'c'(x) (7.15.9)

would indicate that Q was a tensor density of weight WQ. If WQ = 0, then Q is a regular tensor. With this
definition in mind, it is easy to generalize the notion of a "covariant equation" to include tensor densities.
Consider some arbitrary tensor equation which we represent by our example above,

Qadc(x) = Hab(x)Tbc(x) Bd(x) . (7.15.8)

Suppose all four objects Q, H, T, B are tensor densities with weights WQ, WH, WT, WB. If the four objects
Q, H, T, B are tensor densities, and if the up/down free indices match on both sides (the non-contracted

127
 Chapter 7: Standard Notation

indices), and if WQ = WH + WT + WB, then this is a "valid tensor density equation" and covariance is
guaranteed, so it follows that

Q'adc(x') = H'ab(x')T 'bc(x') B'd(x') . (7.15.3)

It is trivial to edit the above proof by just adding weight factors in the right places and then of course they
cancel out on the two sides.

Examples of covariant tensor equations: In special relativity, which happens to involve linear Lorentz
transformations, a fundamental principle is that any "equation of motion" describing anything at all
(particles, EM fields, etc) must be covariant with respect to Lorentz transformations, or it cannot be a
valid equation of motion (ignoring general relativity). An equation of motion must look the same in a
reference frame which is rotated, boosted, or related by any combination of boosts and rotations to some
original frame of reference (see Section 5.14)).
As was noted earlier, the tradition is to write 4-vector indices as Greek letters and 3-vector spatial
indices as Latin letters. For example, we can define the "electromagnetic field-strength tensor" (rank-2)
this way in terms of the 4-vector "vector potential" Aμ:

1
Fμν ≡ ∂μAν - ∂νAμ with Aμ = ( c φ, A) (7.15.10)

where ∂μ means gμα∂α, the contravariant form of the gradient operator. The components are then

(7.15.11)

where c is the speed of light and of course E and B are the electric and magnetic fields. Maxwell's two
inhomogeneous equations (that is, the two with sources) are, in SI units where ε0μ0= 1/c2,

∂νFμν = μ0 Jμ with Jμ = (cρ,J) (7.15.12)

while the two homogeneous equations become

∂αFμν + ∂μFνα + ∂νFαμ = 0 or ∂αFμν + cyclic = 0 . (7.15.13)

Comment: With some effort, one can show that (7.15.12) says curl B - μ0ε0∂tE = μ0J and div E = ρ/ε0 .
Similarly, (7.15.13) says that curl E + ∂tB = 0 and div B = 0, where B = curl A. A is the vector potential,
φ is the scalar potential, J is current density, ρ is charge density, all in SI units.

One can see that each of the above equations involves only tensors and we expect that in x'-space these
equations will take the form

128
 Chapter 7: Standard Notation

1
∂'νF 'μν = μ0 J'μ with J'μ = (cρ',J') and A'μ = ( c φ', A')
∂'αF 'μν + ∂'μF 'να + ∂'νF 'αμ = 0 or ∂'αF 'μν + cyclic = 0 . (7.15.14)

Objects like ∂νFμν and ∂αFμν are true rank-3 tensors because the transformation F is linear. Notice that
both sides of (7.15.12) transform as a contravariant vector, while both sides of (7.15.13) transform as a
covariant rank-3 tensor (the right side is the all-zero rank-3 tensor).

Covariance of tensor equations involving derivatives with non-linear F. A tensor equation which involves
derivatives of tensors is non-covariant under transformations F which are non-linear. The reason is that
the derivative of a tensor is, in that case, not a tensor, as shown in the next Section. Such tensor equations
can be made covariant by replacement of all derivatives by covariant derivatives (which are indicated by
a semicolon). In general relativity, this is known as the Principle of General Covariance (Weinberg p
106). A simple example is the tensor equation gab;c = 0 shown in (F.9.13). Examples relating to the
transformation from Cartesian to curvilinear coordinates appear in Chapter 15.

7.16 The Christoffel Business: covariant derivatives

This subject is treated in full detail in Appendix F, but here we provide some motivation. It should be
noted that a normal derivative is sometimes written ∂aVb = Vb,a with a comma, whereas the covariant
derivative discussed below is written Vb;a with a semicolon.

When a transformation F is non-linear, the matrix Rab is a function of x. Thus one gets the following
transformation for a lower index derivative of a covariant vector field component ∂aVb(x), where a
"second term" quite logically appears,

(∂'aV'b) = (Rad∂d) (RbcVc) = Rad Rbc (∂dVc) + Rad(∂d Rbc)Vc . (7.16.1)

This second term did not arise in (2.4.3) where we looked at ∂a on a scalar field φ'(x') = φ(x) ,

(∂'aφ') = (Rad∂d) φ = Rad (∂d φ). (7.16.2)

In special relativity, for example, where transformations are linear, ∂d Rbc = 0, there is no second term,
and the object ∂aVb transforms as a covariant rank-2 tensor,

(∂'aV'b) = Rad Rbc (∂dVc) . // F is a linear transformation (7.16.3)

But in the general case the second term is present, so ∂dVc fails to transform as a rank-2 covariant tensor.
In this case, one defines a certain "covariant derivative" which itself has an extra piece

Vb;a ≡ ∂aVb – Γkab Vk ⇒ Vd;c ≡ ∂cVd – Γkcd Vk (7.16.4)

where Γcab is a certain function of the metric tensor g. One then finds that

129
 Chapter 7: Standard Notation

V'b;a = Rad Rbc V d;c // the notation ∇cVd ≡ V d;c is also commonly used
or
[∂'aV'b – Γ 'kab V'k] = Rad Rbc [∂dVc – Γkcd Vk] (7.16.5)

so that this covariant derivative of a covariant vector field Vc transforms as a covariant rank-2 tensor
even with non-linear transformation F (see Christoffel Ref., 1869). This issue arises in general relativity
and elsewhere. The object Γcab (sometimes called the "Christoffel connection") is given by

⎧c⎫
Γcab ≡ {ab,c} ≡ ⎨ab⎬ ≡ gcd [ab,d] = ½ gcd( ∂agbd + ∂bgad – ∂dgab ) // Christoffel 2nd kind
⎩ ⎭
(7.16.6)
Γdab ≡ [ab,d] ≡ ½ ( ∂agbd + ∂bgad – ∂dgab ) // Christoffel 1st kind

and this is where the various "Christoffel symbols" come into play. In general relativity and elsewhere,
Γcab is known as the "affine connection" which represents the effect of "curved space" appearing as a
force which acts on a mass (that is to say, a gravitational force), see Section 5.15.
Warning: There is a differently defined version of Γcab floating around in the literature. The version
used above and everywhere in this document is that of Weinberg and is the most common form.
The derivative of any tensor field other than a scalar field shows this same complication when the
underlying transformation F is non-linear. For example, ∂agbd(x) does not transform as a rank-3 tensor,

∂'ag'bd(x') = (Rad∂d)(Rbb'Rdd'gb'd') = RadRbb'Rdd'(∂d gb'd') + other terms (7.16.7)

and therefore neither of the Christoffel symbols Γdab or Γcab transforms as a tensor in this case.
See Appendix F for more detail.

7.17 Expansions of higher order tensors

Appendix E clarifies the use of direct product and polyadic notations for describing the basis vector
combinations onto which higher order tensors can be expanded in a simple generalization of the vector
expansions presented in Section 7.13 above, summarized in (7.13.12). There it was shown that a vector A
can be expanded in two interesting ways :

A = Σn An un An are the contravariant components of A in x-space

A = Σn A'n en . A'n are the contravariant components of A in x'-space (7.17.1)

In the first, un are axis-aligned basis vectors, and in the second en are the tangent base vectors. If A is
instead a tensor of rank n, these expansions are replaced by

A = Σijk... Aijk... (ui⊗uj⊗uk...) Aijk... are the contravariant components of A in x-space

A = Σijk... A'ijk... (ei⊗ej⊗ek...) A'ijk... are the contravariant components of A in x'-space
(7.17.2)

130
 Chapter 7: Standard Notation

where there are n indices in each sum, n factors in the direct products, and n contravariant indices on the
components of tensors A in x-space and in x'-space. In the polyadic notation the direct-product crosses are
eliminated giving

A = Σijk... Aijk... uiujuk... Aijk... are the contravariant components of A in x-space

A = Σijk... A'ijk... eiejek... A'ijk... are the contravariant components of A in x'-space . (7.17.3)

In the case of rank-2 tensors, a product like ui⊗uj = uiuj is called a dyadic (see Section E.4). In this
case (only) the product can be visualized as uiuTj which is a matrix constructed from a column vector to
the left of a row vector. Thus one can write

A = Σij Aij uiuTj Aij are the contravariant components of A in x-space

A = Σij A'ij eieTj . A'ij are the contravariant components of A in x'-space . (7.17.4)

Section E.7 promotes the interpretation of a rank-2 tensor A as an operator in a Hilbert space, where the
matrices Aij and A'ij are matrices associated with the operator A in different bases,

Anm = <un | A | um > = the x-space components of tensor A

A'nm = <en | A | em > = the x'-space components of tensor A . (7.17.5)

These expansion methods are used in Appendices G and H to derive curvilinear expressions for two
objects that play a role in continuum mechanics, (∇v) and div(T) (where T is a tensor).

7.18 Collection of Facts about basis vectors en , u'n and bn.

Recall this Developmental Notation block of information regarding the en family of basis vectors, (6.4.1),

( ē n)i = ḡ ij (en)j (Ēn)i = ḡ ij (En)j [ ē n = ḡ en Ēn = ḡ En] (5.8.4)

(e'n)i = Rij(en)j (E'n)i = Rij(En)j [ e'n = R en E'n = R En ]

( ē 'n)i = Sji( ē n)j (Ē'n)i = Sji(Ēn)j [ ē 'n = ST ē n Ē'n = ST Ēn ] (2.5.1)

(en)i = Sin (En)i = gijRnj = g'njSij (en')i = δi,n (En')i = g'ni

( ē n)i = ḡ ijSjn = Rji ḡ'jn (Ēn)i = Rni ( ē n')i = ḡ'ni (Ēn')i = δn,i
(6.3.3) (6.3.9)
en • em = ḡ'nm ⇒ |en| = ḡ'nn = h'n (scale factor) En = g'ni ei
En • em = δn,m en = ḡ'ni Ei
En • Em = g'nm ⇒ |En| = g'nn . (6.2.4)

131
 Chapter 7: Standard Notation

e'n • e'm = ḡ'nm ⇒ |e'n| = ḡ'nn = h'n (scale factor) E'n = g'ni e'i
E'n • e'm = δn,m e'n = ḡ'ni E'i
E'n • E'm = g'nm ⇒ |E'n| = g'nn . (6.2.7)

(Ēn)i(en)j = δi,j [ Σn Ēn enT = 1 ] (6.2.16) and (6.2.24) DN (6.4.1)

We shall now manually translate this entire block into Standard Notation. In doing so, we use

gab → gab // the contravariant rank-2 metric tensor g (and similarly for g')
ḡ ab → gab // the covariant rank-2 metric tensor ḡ (7.4.1)

Rij → Rij (7.5.2)

Sik → Sik (7.5.4)

Sij = Rji (7.5.13)

and (7.13.1) for everything else. Matrix equations are ignored and only their component forms are
translated. Here is the result,

(en)i = gij (en)j (en)i = gij (en)j (5.8.4)

(e'n)i = Rij(en)j (e'n)i = Rij(en)j

(e'n)i = Sji(en)j = Rij(en)j (e'n)i = Sji(en)j = Rij(en)j (2.5.1)

(en)i = Sin = Rni (en)i = gijRnj = g'njSij (e'n)i = δni (e'n)i = g'ni
(en)i = gijSjn = Rjig'jn (en)i = Rni (e'n)i = g'ni (e'n)i = δni
(6.3.3) (6.3.9)
en • em = g'nm ⇒ |en| = g'nn = h'n (scale factor) en = g'ni ei
en • em = δnm en = g'ni ei
en • em = g'nm ⇒ |en| = g'nn . (6.2.4)

e'n • e'm = g'nm ⇒ |e'n| = g'nn = h'n (scale factor) e'n = g'ni e'i
e'n • e'm = δnm e'n = g'ni e'i
e'n • e'm = g'nm ⇒ |e'n| = g'nn . (6.2.7)

(en)i(en)j = δij (6.2.16) // completeness SN (7.18.1)

Recall that {En, en} are a dual pair in the sense of Section 6.2, For the inverse transformation, we also had
{Un', u'n} as a dual pair in Section 6.5, and the block of data about the u'n family of vectors was stated
in (6.5.3) in developmental notation. We could translate that into standard notation as we did the above,
but it is easier to translate (7.18.1) using the rules (6.5.2)

g'↔ g R↔S en → u'n e'n → un En → U'n E'n → Un (6.5.2)

132
 Chapter 7: Standard Notation

translated to Standard Notation

g'↔ g R↔S en → u'n e'n → un en → u'n e'n → un (7.18.2)

Applying these rules one at a time to (7.18.1) we obtain this data block for the u'n family,

(u'n)i = g'ij (u'n)j (u'n)i = g'ij (u'n)j (5.8.4)

(un)i = Sij(u'n)j (un)i = Sij(u'n)j

(un)i = Rji(u'n)j = Sij(u'n)j (un)i = Rji(u'n)j = Sij(u'n)j (2.5.1)

(u'n)i = Rin = Sni (u'n)i = g'ijSnj = gnjRij (un)i = δni (un)i = gni
(u'n)i = g'ijRjn = Sjigjn (u'n)i = Sni= Rin (un)i = gni (un)i = δni
(6.3.3) (6.3.9)
u'n • u'm = gnm ⇒ |u'n| = gnn = hn (scale factor) u'n = gni u'i
u'n • u'm = δnm u'n = gni u'i
u'n • u'm = gnm ⇒ |u'n| = gnn . (6.2.4)

un • um = gnm ⇒ |un| = gnn = hn (scale factor) un = gni ui

un • um = δnm un = gni un
un • um = gnm ⇒ |un| = gnn . (6.2.7)

(u'n)i(u'n)j = δij (6.2.16) // completeness SN (7.18.3)

Finally, we shall restate (7.18.1) for the generic dual pair {Bn, bn} of Section 6.2. Without detailed proof,
we claim that one can in general construct a transformation x' = Fb(x) for which some generic complete
set of vectors bn(x) are the tangent base vectors like en(x) are for x' = F(x). For this transformation Fb,
one would have (bn)i = (Sb(x))in from (3.2.6), so matrix Sb(x) would be known. Then Rb = (Sb)-1 is also
known, and one would then integrate (2.1.6) dx' = Rb(x) dx to obtain a viable Fb(x).
As discussed in Section 6.2, for this given set of {Bn, bn}, we defined W'nk ≡ bn•bk in developmental
notation which is analogous to ḡ'nk = en • ek for F(x). Thus suggests we rename W'→w̄' and then we
have ḡ'nk = en • ek for F(x) and w̄'nk = bn • bk for Fb(x), where w̄' = W. In Standard Notation these
equations become g'nk = en • ek and w'nk = bn • bk since

W'nk = w̄'nk → w'nk like ḡ'nk → g'nk

w'nk → w'nk g'nk → g'nk . (7.18.4)

So the rules for translating (7.18.1) from en to generic bn are are very simple :

133
 Chapter 7: Standard Notation

e→b // that is, en → bn, en → bn and the same for primed basis vectors
g' → w' // that is, g'ij → w'ij and g'ij→ w'ij (7.18.5)

Note that the arrows → in (7.18.4) are for DN→SN, whereas those in (7.18.5) are SN e → SN b.

The result is

(bn)i = gij (bn)j (bn)i = gij (bn)j (5.8.4)

(b'n)i = Rij(bn)j (b'n)i = Rij(bn)j

(b'n)i = Sji(bn)j = Rij(bn)j (b'n)i = Sji(bn)j = Rij(bn)j (2.5.1)

(bn)i = Sin = Rni (bn)i = gijRnj = w'njSij (bn')i = δni (b'n)i = w'ni
(bn)i = gijSjn = Rjiw'jn (bn)i = Rni (bn')i = w'ni (b'n)i = δni
(6.3.3) (6.3.9)
bn • bm = w'nm ⇒ |bn| = w'nn = h'n (scale factor) bn = w'ni bi
bn • bm = δnm bn = w'ni bi
bn • bm = w'nm ⇒ |bn| = w'nn . (6.2.4)

b'n • b'm = w'nm ⇒ |b'n| = w'nn = h'n (scale factor) b'n = w'ni b'i
b'n • b'm = δnm b'n = w'ni b'i
b'n • b'm = w'nm ⇒ |b'n| = w'nn . (6.2.7)

(bn)i(bn)j = δij (6.2.16) // completeness SN (7.18.6)

7.19 More on basis vectors and matrix elements of R and S

Ambiguity

In the summary tables of Section 7.18 we see various equations like these,

(un)i = δni (en)i = Rni (e'n)i = δni (u'n)i = Rin . (7.19.1)

There is a certain ambiguity in writing, for example, (en)i : it is not clearly stated in what basis
component i is evaluated! The basis depends on the expansion in which a component appears. For
example, from (7.13.10) we write for a generic vector V,

V = ΣiVi ui ui • V = Vi (7.13.10)

and we then know that Vi is a component in the u-basis. Applying this to V = en (an x-space vector) we
conclude that

134
 Chapter 7: Standard Notation

en = Σi (en)i ui ui • en = (en)i . (7.19.2)

Thus, (en)i is in fact a component in the u-basis, not for example the e-basis. We would then write

(en)i = (en)(u)i = Rni . (7.19.3)

This is the "natural basis" to use for the components of en and for any x-space vector V where we use the
natural expansion V = ΣiVi ui shown above.

One could however expand en on the en basis to get

en = Σi (en)(e)i ei ei • en = (en)(e)i = δin . (7.19.4)

Then with no ambiguity one could write

(en)(u)i = Rni // natural

(en)(e)i = δni (7.19.5)

Thus, when we write (en)i, we imply that we are using the u-basis for the components. Since this applies
to any x-space vector, it is also true for (un)i, so (un)i = (un)(u)i = ui • un .

What about x'-space vectors? From (7.13.11) we write

V' = Σi(V')i e'i e'i • V' = (V')i . (7.13.11)

Applying this to V' = e'n (an x'-space vector) we conclude that

e'n = Σi(e'n)i e'i e'i • e'n = (e'n)i = δin . (7.19.6)

Thus, (e'n)i is in fact a component in the e'-basis, not for example the u'-basis. We would then write

(e'n)i = (e'n)(e')i = δni . (7.19.7)

This is the "natural basis" to use for the components of e'n and for any x'-space vector V' where we use
the natural expansion V' = Σi(V')i e'i shown above.

One could however expand e'n on the u'n basis to get

e'n = Σi (e'n)(u')i u'i u'i • e'n = (e'n)(u')i = ui • en = (en)i = Rni . (7.19.8)

Then with no ambiguity one could write

(e'n)(e')i = δni // natural

(e'n)(u')i = Rni (7.19.9)

135
 Chapter 7: Standard Notation

Thus, when we write (e'n)i, we imply that we are using the e'-basis for the components. Since this applies
to any x'-space vector, it is also true for (u'n)i, so (u'n)i = (u'n)(e')i = e'i • u'n.

We summarize the above discussion as follows:

Fact: un and en components are by default presented in the u-basis (7.19.10)

e'n and u'n components are by default presented in the e'-basis

Table of basis vector dot products

Since there are four kinds of basis vectors of interest, un, en, u'n and e'n with lower labels, one might
imagine there are 16 scalar products of interest,

u e u' e'
u u•u u•e u•u' u•e'
e e•u e•e e•u' e•e'
u' u'•u u'•e u'•u' u'•e'
e' e'•u e'•e e'•u' e'•e'

However, since a scalar product only exists within a Hilbert space (x-space or x'-space), the cross space
entries in this table make no sense, so we eliminate them to get

u e u' e'
u u•u u•e - -
e e•u e•e - -
u' - - u'•u' u'•e'
e' - - e'•u' e'•e' (7.19.11)

Since u•e = e•u and u'•e' = e'•u', we see that there are only 6 distinct scalar products of interest.
Collecting data from the tables (7.13.1) and (7.13.2) we summarize these 6 cases as follows:

Table of basis vector dot products:

(un)i = ui • un = <ui | un > = gin = u'i • u'n = < u'i | u'n >

(en)i = ui • en = <ui | en > = Sin = Rni

(en')i = e'i • e'n = < e'i | e'n > = g'in = ei • en = <ei | en >

(un')i = e'i • u'n = <e'i | u'n> = Rin = Sni (7.19.12)

136
 Chapter 7: Standard Notation

Since a • b = a' • b', the first and third lines each contain 2 of the 6 cases. We are careful to use the
correct tensor notation gni and g'ni for δni as discussed near equation (7.4.17). Doing this, we may state
the following :

Fact: In any of the above equation lines, one may raise the label n and/or lower the label/index i and end
up with another valid equation line. (7.19.13)

In this manner, from the 4 lines above one may generate 12 more lines of equations.

Note that we have slipped in the Dirac notation < | > described in (E.7.4) as an alternate way to write the
dot products.

As an example of the above Fact, if we raise the label n in the third line we get these valid equations,

(en')i = e'i • en' = < e'i | en' > = g'in = ei • en = <ei | en >

and if we then lower index i, we get another set of valid equations,

(en')i = e'i • en' = < e'i | en' > = g'in = ei • en = <ei | en > .

To prove the above Fact, one first observes in which space a vector lies, so one knows which metric
tensor raises and lowers the index. Then one uses these facts from (7.18.1) and (7.18.3),

en = g'ni ei e'n = g'ni e'i un = gni ui u'n = gni u'i

en = g'ni ei e'n = g'ni e'i un = gni un u'n = gni u'i . (7.19.14)

For example, consider the third line in (7.19.12),

(en)i = ui • en = <ui | en > = Rni .

Apply g'mn to get

g'mn (en)i = g'mn ui • en = g'mn <ui | en > = g'mn Rni

or
[g'mn en]i = ui • [g'mn en] = <ui | [g'mn] en > = g'mn Rni
or
(em)i = ui • [em] = <ui | em > = Rmi // see (7.5.9) about Rmi

which shows that the label n can be raised to get another valid set of equations. Now start with the same
third line in (7.19.12) and instead apply gji to get

gji(en)i = gjiui • en = gji<ui | en > = gjiRni

or
gji(en)i = [gjiui] • en = <gjiui | en > = gjiRni
or
(en)j = uj • en = <uj | en > = Rnj // see (7.5.9) about Rnj

137
 Chapter 7: Standard Notation

which shows that the index i can be lowered to get another valid set of equations.

Matrix elements of R and S

One can regard the elements Rij and Sij as being certain matrix elements of operators R and S, and this
is most easily handled in the Dirac bra-ket notation of (E.7.4). Since V' = RV (vector transformation rule)
one sees that a matrix element <a' | R | b> = <a' | R b> must have | b> being an x-space vector, and <a' |
being a transposed x'-space vector.

We shall now compute matrix elements of R in several basis combinations. As examples of V' = RV we
know that,

e'n = R en
u'n = R un
or
|e'n> = R |en>
|u'n> = R |un> . (7.19.15)

We can "close on the left" in various ways. For example,

<e'i | e'n> = <e'i | R | en> = g'in

<u'i | e'n> = <u'i | R | en> = Sin = Rni (7.19.16)

Here we look up the scalar products on the left of each line using table (7.19.12), making use of Fact
(7.19.13), and adjusting the various index names and the up and down index sense.

Now start instead with the second equation of (7.19.15) and close two ways, again looking up the basis
vector scalar products in (7.19.12),

<e'i |u'n> = <e'i | R | un> = Rin = Sni

<u'i |u'n> = <u'i | R | un> = gin . (7.19.17)

Here then are four R matrix elements of interest,

<e'i | R | en> = g'in

<u'i | R | en> = Sin = Rni

<e'i | R | un> = Rin = Sni

<u'i | R | un> = gin . (7.19.18)

138
 Chapter 7: Standard Notation

To get matrix elements of S, we use the fact (7.9.8) that R = ST and the fact (7.9.17) that <a|S|b> =
<b|ST|a> = <b|R|a>, so we translate the above four lines:

<en | S | e'i> = <e'i | R | en> = g'in

<en | S | u'i> = <u'i | R | en> = Sin = Rni

<un | S | e'i> = <e'i | R | un> = Rin = Sni

<un | S | u'i> = <u'i | R | un> = gin . (7.19.19)

and the above then is a full statement of all matrix elements of S and R, for a particular up down sense of
the basis vector labels. As usual, g'in = gin = δin but we maintain the true tensor form to allow for up
down modifications as per the following claim:

Fact: In the above equations, one can raise/lower labels on either or both sides to get new valid
equations. (7.19.20)

The proof of this claim is the same as the proof of Fact (7.19.13) above using (7.19.14). For example,

<u'i | R | un> = gin

so
gji<u'i | R | un> = gjigin
or
<u'j | R | un> = gjn .

Similarly <e'i | R | un> = Rin = Sni ⇒ <e'i | R | un> = Rin = Sni.

Interpretation of Rin and Sin

Since V' = R V, one can regard the operator R as a mapping R : X → X' (x-space → x'-space). R may be
regarded as a "cross tensor" having one foot in each space. As shown in (7.19.19),

Rij = <e'i | R | uj> = [ R(e',u)]ij (7.19.21)

so one can regard R as an abstract cross-tensor being expanded as follows in the mixed-basis sense of
(E.10.4),

R = Σij [ R(e',u)]ij (e'i ⊗ uj) . (7.19.22)

As usual, the coefficients (the cross tensor components) can be projected out using

139
 Chapter 7: Standard Notation

(e'a ⊗ ub) • R = Σij [ R(e',u)]ij (e'a ⊗ ub) • (e'i ⊗ uj)

= Σij [ R(e',u)]ij (e'a • e'i) ⊗ (ub • uj)

= Σij [ R(e',u)]ij δai δbj

= [ R(e',u)]ab . (7.19.23)

Similarly, since V = S V' one can regard the operator S as a mapping S : X' → X where, from (7.19.19),

Sij = <ui | S | e'j> = [ S(u,e')]ij (7.19.24)

and then the cross-tensor expansion and projection is given by,

S = Σij [ S(u,e')]ij (ui ⊗ e'j)

(ua ⊗ e'b) • S = [ S(u,e')]ab . (7.19.25)

A Pitfall Example

One knows from (7.19.19) that

< u'i | R ua> = < u'i | R | ua> = gia = δia .

This is also known from,

< u'i | R ua> = < u'i | u'a> = δia .

Consider now the following slight of hand wherein we get a different result,

< u'i | R ua> = [R ua]i = Rij(ua)j = Rijδaj = Ria wrong!

Where is the error being made here? Well, the vector u'a = Rua being in x'-space has a natural e'-type
basis as discussed below (7.19.7), but here the component [Rua]i is in the u'-basis. When doing
something "unnaturally" one must pay more attention to labels. The correct version of the above is

< u'i | R ua> = [R ua](u')i = [R(u',u)]ij (ua)(u)j = [R(u',u)]ij δaj

= [R(u',u)]ia = <u'i | R | ua> = gia = δia . (7.19.26)

Now consider instead a different example. From (7.19.19) one knows that,

< e'i | R ua> = < e'i | R | ua> = Ria .

In this case write

140
 Chapter 7: Standard Notation

< e'i | R ua> = [R ua]i = Rijδaj = Ria

and there is no problem because e' is the natural basis for x'-space. In more detail

< e'i | R ua> = [R ua](e')i = [R(e',u)]ij (ua)(u)j = [R(e',u)]ij δaj

= [R(e',u)]ia = <e'i | R | ua> = Ria . (7.19.27)

We find that the Dirac notation is useful because it provides a bulletproof formalism for avoiding
ambiguities such as that of the previous example. The above equation sequence can be written (implied
sum on j as usual)

< e'i | R ua> = < e'i | R [1] ua> = < e'i | R [ | uj><uj | ] ua>

= < e'i | R | uj><uj | ua> = [R(e',u)]ij (ua)(u)j (7.19.28)

where we use the (E.7.4) completeness relation 1 = | uj><uj |. In fact

x-space completeness: 1 = | uj><uj| = | uj><uj| = | ej><ej| = | ej><ej|

x'-space completeness: 1 = | u'j><u'j| = | u'j><u'j| = | e'j><e'j| = | e'j><e'j| (7.19.29)

There are just restatements of the (6.2.8) duality idea that bi • bj = < bi | bj > = δij for any basis.

141
 Chapter 8: Length Area and Volume

8. Transformation of Differential Length, Area and Volume

8.1 Overview of Chapter 8

This Chapter and all remaining Chapters use the Standard Notation introduced in Chapter 7.
The term N-piped is short for N dimensional parallelepiped.
The context is Picture B:

(8.1.1)

Since this Chapter is quite lengthy, a brief overview is in order:

The transformation of differential length, area, and volume is first framed in terms of the mapping of an
orthogonal differential N-piped in x'-space to a skewed differential N-piped in x-space. The N-piped in x'-
space has axis-aligned edges of length dx'n, while the N-piped in x-space has edges endx'n where en are
the tangent base vectors introduced in Chapter 3. We want to learn what happens to the edges, face areas
and volume as one differential N-piped is mapped into the other by the curvilinear coordinate
transformation x' = F(x). After solving this problem, we go on to consider the transformations of arbitrary
differential vectors, areas and volume.

8.2: The differential N-piped mapping

The differential N-piped mapping is described and various symbols are defined.

8.3: Properties of the finite N-piped spanned by the en in x-space

Results from Appendix B concerning finite N-piped geometric properties are quoted. Certain definition
changes are made to make the formulas suitable for tensor analysis. The purpose of the lengthy Appendix
B is to lend credence to the general formulas for elements of area and volume in N dimensions.

8.4: Back to the differential N-piped mapping: how edges, areas and volume transform
(a) Setup. The finite N-piped edges en are scaled by curvilinear coordinate variations dx'n to create a
differential N-piped in x-space having edges (endx'n).
(b) Edge Transformation. The edges en of the x-space N-piped map to axis-aligned edges e'n in x'-space.
(c) Area Transformation. Tensor density methods as presented in Appendix D are used here.
(d) Volume Transformation. The volume transformation is computed several different ways.
(e) Covariant Magnitudes. These are | dx'(n)|, | dA'n | and | dV' | in the Curvilinear View of x'-space.
(f) Two Theorems. (1) g'nn g' = cof(g'nn) and (2) |(Πxi≠nei)| = cof(g'nn) .
(g) Cartesian-View Magnitude Ratios. Appropriate for the continuum mechanics application.
(h) Nested Cofactor Formulas and STS notation: Description of all "leaf" areas of an N-piped in both
descending and ascending orders.

142
 Chapter 8: Length Area and Volume

(i) Transformation of arbitrary differential vectors, areas and volume. Having built confidence with the
general form of vector area and volume expressions in the N-piped case, the N-piped is jettisoned and
formulas for the transformation of arbitrary vectors, areas and volume are derived.
(j) Concatenation of Transformations. What happens to the transformation of vectors, areas and volumes
when two transformations are concatenated? One result is that J = J1J2.
(k) Examples of area magnitude transformation for N = 2,3,4
Example 2: Spherical Coordinates: area patches

8.5: Transformation of Differential Volume applied to Integration

The volume transformation obtained in Section 8.4 is related to the traditional notion of the Jacobian
changing the "measure" in an integration. The "Jacobian Integration Rule" can then be expressed as a
distributional equation.

8.6: Interpretations of the Jacobian

8.7: Volume integration of a tensor field under linear transformations

Under suitable conditions, the volume integration of a tensor field integrand yields a tensor of the same
type.

8.2 The differential N-piped mapping

Consider these relations involving a differential distance in the n-direction (dx'n > 0):

dx'(n) = e'n dx'n x'-space axis-aligned differential vector, and (e'n)i = δni (3.2.2)

dx(n) = en dx'n x-space mapping of the above vector under F-1 or R-1

dx'(n) = R(x) dx(n) relation of the two differential vectors (contravariant rule) (2.1.6)

e'n = R(x) en (3.3.2) (8.2.1)

A superscript (n) on the differentials makes clear there is no implied sum on n. Note that application of
R(x) to the second equation yields the first equation, making use of the last two equations, so the second
equation really does contain dx'n. In more generality, recall V = Σn V'n en from (7.13.10).
The vectors dx(n) span a differential N-piped in x-space, while the dx'(n) span a corresponding
differential N-piped in x'-space. The two N-pipeds are related by the mapping x' = F(x). Since the regions
are differentially small, this mapping is the same as the linearized mapping dx' = R dx.
The metric tensor in x-space will be taken to be g = 1, so it is a Cartesian space.
As discussed at the end of Section 5.16, the x'-space N-piped can be viewed in (at least) two ways
depending on how the metric tensor g' is set. For a continuum mechanics flow application, one sets g' = 1
and this gives the Cartesian View of the x'-space N-piped. For such flows dot products and magnitudes of
vectors like dx(n) are not invariant under the transformation. For our curvilinear coordinates application,
however, we set g' = RgRT = RRT and this causes vector dot products and magnitudes to be invariant and
we can talk about such objects as being tensorial scalars. This is the Curvilinear View of x'-space.

143
 Chapter 8: Length Area and Volume

When g' ≠ 1, it is impossible to accurately represent the Curvilinear-View picture of the x'-space N-
piped as a drawing in physical space (for N=3). This subject is discussed for a sample 2D system in
Section C.5. Although the basis vectors (e'n)i = δni in x'-space are always axis-aligned, they are only
orthogonal for an orthogonal coordinate system, since e'n•e'm = en•em = g'nm . Nevertheless, even for a
non-orthogonal system we draw the axes as if they were orthogonal, which at least provides a
representation of the notion of "axis aligned" basis vectors. For N > 3 one at least imagine this kind of
drawing.
Due to these graphical difficulties, in the drawing below the Cartesian View of x'-space is shown.
Since g' = 1 for this situation, the axis-aligned basis vectors e'n are in fact unit vectors ^e 'n and are
orthogonal, so the picture becomes at least comprehensible:

(8.2.2)

The orthogonal Cartesian-View N-piped allows visualization of these curvilinear coordinate variations, all
dx'k > 0 ( objects below are primed because they exist in x'-space),

dL'n ≡ dx'n length n = 1,2....N

dA'n ≡ Πi≠ndx'i area
dV' ≡ Πidx'i = dA'n dL'n . volume (no implied sum on n) (8.2.3)

For example, for N=3 one would have

dL'1 ≡ dx'1 dL'2 ≡ dx'2 dL'3 ≡ dx'3

1 2 3 2 3 1
dA' = dx' dx' dA' = dx' dx' dA'3 = dx'1dx'2
dV' = dx'1dx'2dx'3 = dA'1dL'1 = dA'2dL'2 = dA'3dL'3 . (8.2.4)

The Cartesian-View x'-space N-piped is always orthogonal because the (e'n) are orthonormal axis-aligned
unit vectors (since g'=1). In contrast, the x-space N-piped is typically rotated and possibly skewed as well
(if the coordinates x'i describe a non-orthogonal coordinate system). The transformation F and its

144
 Chapter 8: Length Area and Volume

linearized version R map the skewed x-space N-piped into the orthogonal x'-space N-piped. As one
moves around in x-space so that point x changes, the picture on the left above keeps its shape, just
translating itself to the new point x', but the picture on the right changes shape and volume because the
vectors en(x) are functions of x.
It is our goal to write expressions for edges, areas and volumes in these two spaces and to then show
how these objects transform between the two spaces. To this end, we shall rely on work done in Appendix
B which is summarized in the next Section. Following that, we shall add to each edge a differential
distance associated with that edge (such as er → erdr in spherical coordinates), and that will bring us
back to the differential N-piped picture above.

8.3 Properties of the finite N-piped spanned by the en in x-space

The finite N-piped spanned by the tangent base vectors en in x-space has the following properties (as
shown in Appendix B) :

• The N spanning edges are the vectors en which have lengths |en| = h'n (scale factors, (5.11.7) ). (8.3.1)

• There are 2N vertices. (8.3.2)

• There are N pairs of faces. The two faces of each pair are parallel in N dimensions. One face of each
pair touches the point where the tails of all the en vectors meet (the near face) while the other does not
touch this meeting point (the far face). (8.3.3)

• Each face of an N-piped is an (N-1)-piped having 2N-1 vertices. The faces are planar surfaces of
dimension N-1 embedded in an N dimensional space. (8.3.4)

• A face's vector area An is spanned by all the ei except en and is labeled by this missing en vector.
(8.3.5)
n n
• The far face has out-facing vector area A , and the near face has out-facing area vector – A . These
vector areas are normal to the faces. (8.3.6)

• The vector area An is given by several equivalent expressions:

An = |det(Sab)| en

An = σ (-1)n-1 Πxi≠n ei

An = σ (-1)n-1 e1 x e2 ... x eN // en missing σ ≡ sign[det(Sab)] = sign[det(Rab)]

(An)i = σ (-1)n-1 εiabc..x (e1)a(e2)b.... (eN)x // en missing (8.3.7)

• The volume of the N-piped is given by (see Section 5.12 concerning J)

V = | det [e1, e2, e3 ... eN] | = | det(Sab) | = g'1/2 = |J| . (8.3.8)

145
 Chapter 8: Length Area and Volume

The mapping picture above does not apply to a finite N-piped. The finite N-piped just discussed exists in
x-space. One might ponder into what shape it maps in x'-space under the transformation F. In the case of
spherical coordinates, Fig (1.13), all of x-space maps into a certain orthogonal "office building" in x'-
space. A finite N-piped within x-space maps into some very complicated 3D shape within the office
building which is bounded by curves which in general are not even coplanar. The point is that a finite N-
piped in x-space does not map back into some nice orthogonal N-piped in x'-space and the picture drawn
in the previous Section does not apply. However, when differentials are added in the next Section, then,
since the mapped regions are very small, the mapping of the x-space differential N-piped is in fact an
"orthogonal" N-piped in x'-space. This is because for a tiny region near some point x, the mapping
between dx and dx' is described by the linear relation dx' = R(x) dx of (2.1.6).

Conventions for defining the area vector and volume. In Appendix B (as reported in (8.3.6)) the area
vector An is defined so that the out-facing normal of the x-space N-piped's "far face n" is An, regardless
of the sign of det(S). This was done to simplify the computation of the flux of a vector field emerging
from the N-piped in the geometric divergence calculation in Chapter 9. That calculation is then valid for
either sign of det(S), a sign that we call σ. The following picture illustrates on the left an x-space 3-piped
which is obtained by reverse-mapping the x'-space orthogonal 3-piped using an S which has det(S) > 0.
On the right is the x-space 3-piped that results for S → -S. These two 3-pipeds are related by a parity
inversion of all points through the origin. If the origin lies far away, these two N-pipeds lie far away from
each other, a fact not illustrated in the picture:

(8.3.9)
3
Notice that A for the "far face 3" is outfacing in both cases.
This definition of the vector area is not suitable for the vector analysis we are about to undertake.
Instead of the above situation, we will redefine A3 = e1 x e2 for both pictures, and this will cause the A3
vector in the right picture to point up into the interior of the N-piped. This new definition allows us to
interpret A3 = e1 x e2 as a "valid vector equation" to which we may apply the ideas of covariance and
tensor densities.
Notice that under a parity transformation, this newly defined A3 is a "pseudovector" which is one
which does not reverse direction under a parity transformation, since A3 = (-e1) x (-e2). The subject of
parity and handedness and the sign of det(S) is discussed more in Section 6.9.

146
 Chapter 8: Length Area and Volume

Here then are the expressions for An with this new definition, where the new forms are obtained from
(8.3.7) by multiplying by σ = sign ( det(S) ) :

An = det(Sab) en = J en (8.3.10)

An = (-1)n-1 e1 x e2 ... x eN // en missing (8.3.11)

(An)i = (-1)n-1 εiabc..x (e1)a(e2)b.... (eN)x // en missing (8.3.12)

The second line shows that pseudovector areas can only exist for an odd number of dimensions (such as
N=3).
A similar redefinition of the volume will now be done. In Appendix B the volume is defined so as to
be a positive number regardless of σ with the result V = |det(S)|. We now redefine the volume by
multiplication by σ, so that now V = det(S) which is of course a negative number when det(S) < 0, which
in turn means the en are forming a left-handed coordinate system as per Section 6.9. So:

V = det [ e1, e2, e3 ... eN] = det(Sab) = J . (8.3.13)

8.4 Back to the differential N-piped mapping: how edges, areas and volume transform

(a) The Setup

If the edges of the finite N-piped described above are scaled by positive differentials dx'n > 0, the result is
a differential N-piped in x-space which has these properties:

dx(n) = en dx'n (8.2.1) // edges

dAn = J en (Πi≠ndx'i) (8.3.10) // areas

dAn = (-1)n-1 (dx'1e1) x (dx'2e2) ... x (dx'NeN) (8.3.11) // en missing from cross product

= (-1)n-1 e1 x e2 ... x eN (Πi≠ndx'i) // en missing from cross product

= (-1)n-1 (Πxi≠nei) (Πi≠ndx'i) // shorthand notation of (A.10.5)

(dAn)i = (-1)n-1 εiabc..x (e1)a(e2)b.... (eN)x (Πi≠ndx'i) // en factor and index missing, (8.3.12)

dV = det [dx'e1, dx'e2, dx'e3 ... dx'eN] = det [ e1, e2, e3 ... eN] (Πidx'i) (8.3.13)

= det(Sab)(Πidx'i) = J (Πidx'i) // (7.13.2) and (5.12.6)

= εabc..x (dx'e1)a(dx'e2)b..... (dx'eN)x // (5.12.8)

(8.4.a.1)

147
 Chapter 8: Length Area and Volume

where dAn and dV are obtained from An and V according to the new definitions described above. These
equations apply to the right side of the N-piped mapping picture Fig (8.2.2) which is replicated here :

(8.2.2)

For spherical coordinates, the N-piped on the right would be spanned by these vectors, see (3.4.6),

erdr = ^
r dr, eθdθ= r^
θ dθ ^dφ .
eφdφ = rsinθ φ (8.4.a.2)

The reader is reminded that the vectors shown on the right of (8.2.2) exist in x-space, and that this x-space
is assumed to be Cartesian for this entire Chapter, meaning g = 1 and Vk= Vk for vectors.

(b) Edge Transformation

The edge dx(n) we know from (2.1.6) transforms as a tensorial vector under transformation F, so

dx'(n) = R dx(n) where dx(n) = en dx'n (8.2.1) . (8.4.b.1)

Making use of (7.18.1) and RS = 1, evaluation gives

[dx'(n)]i = Rij (en)j dx'n = Rij Sjn dx'n = (RS)in dx'n = δin dx'n

⇒ dx'(n) = e'n dx'n since (e'n)i = δni (8.4.b.2)

so this contravariant edge points along the the n axis direction in x'-space of Fig (8.2.2). This fact was
stated above in (8.2.1), we are just exercising our notation.

148
 Chapter 8: Length Area and Volume

(c) Area Transformation

How does dAn transform under F? Looking at the component form stated above in (8.4.a.1),

(dAn)i = (-1)n-1 εiabc..x (e1)a(e2)b.... (eN)x (Πi≠ndx'i) // en factor and index missing (8.4.c.1)

it is seen that dAn is a combination of tensor objects like εiabc..x and (e2)b. As discussed in Appendix
D, a vector density of weight 0 is an ordinary vector such as e2 or dx. The ε tensor (rank-N) is a tensor
density of weight -1. In forming more complicated tensor objects, the rule (D.2.3) is that one adds the
weights of the objects being combined. Therefore, one may conclude that the object dAn is a vector
density of weight -1. This has the immediate implication that dAn transforms to x'-space under F
according to the rule (D.1.4),

dA'n = J R dAn or (dA'n)i = J Rij (dAn)j // J-W = J-(-1) = J (8.4.c.2)

where J = det(S) is the Jacobian of Section 5.12. Moreover, the above equation for (dAn)i is a "valid
tensor density equation" as per (7.15.9) and is therefore covariant. This means that in x'-space the
equation has the exact same form, but tensor objects are primed (the dx'i are constants),

(dA'n)i = (-1)n-1 ε'iabc..x (e'1)a(e'2)b.... (e'N)x (Πi≠ndx'i) // e'n factor and index missing
(8.4.c.3)
Insertion of (D.5.13) that ε'iabc..x = J2 εiabc..x and (e'n)i = δni then gives

(dA'n)i = (-1)n-1 J2 εiabc..x δ1aδ2b...δNx (Πi≠ndx'i)

= (-1)n-1 J2 εi123..N (Πi≠ndx'i) // index n missing on ε

= δni J2 (Πi≠ndx'i) (8.4.c.4)

The last step follows from the fact that εi123..N with n missing must vanish if i ≠n, and if i=n then
εn123..N = (-1)n ε123..n..N = (-1)n. The conclusion then is that

dA'n = J2 (Πi≠ndx'i) e'n since (e'n)i = δni (7.18.1) (8.4.c.5)

and the covariant vector area dA'n points in the n-axis direction in x'-space.
For the covariance-dubious reader, here is an alternate derivation of (8.4.c.4) from (8.4.c.2) :

(dA'n)i = J Rij (dAn)j = J Rij {[ (-1)n-1 εiabc..x (e1)a(e2)b.... (eN)x (Πi≠ndx'i) } // en missing

= J Rij {[ (-1)n-1 εiabc..x R1a R2b.... RNx (Πi≠ndx'i) } // Rnκ missing

= J {[ (-1)n-1 εiabc..x Rij R1a R2b.... RNx (Πi≠ndx'i) } // Rnκ missing

= J {[ (-1)n-1 εiabc..x Sji Sa1 Sb2.... SxN (Πi≠ndx'i) } . // Sκn missing (8.4.c.6)

149
 Chapter 8: Length Area and Volume

where Rij = Sji by (7.5.13). The εSSS..S object is the determinant of an N-1 dimensional matrix. If i ≠n,
then i must be some index like 1 or 2. But then the determinant has two identical columns, so vanishes.
Thus, the result is proportional to δni. Continuing,

= δni J { (-1)n-1 εnabc..x Sjn Sa1 Sb2.... SxN (Πi≠ndx'i) } // Sκn missing in group

= δni J { (-1)n-1 εnabc..x Sa1 Sb2.... Sjn.... SxN (Πi≠ndx'i) }

= δni J { εabc..n...x Sa1 Sb2.... Sjn.... SxN (Πi≠ndx'i) }

= δni J { det(S) (Πi≠ndx'i) }

= δni J2 (Πi≠ndx'i) (8.4.c.7)

which agrees with the result (8.4.c.4) from the covariant x'-space equation.

(d) Volume Transformation

What about the volume dV? Assume for the moment that dV is correctly represented this way, where any
n will do:

dV = dAn • dx(n) = (dAn)i [dx(n)]i . // no implied sum on n (8.4.d.1)

Installing our (8.4.a.1) expressions for dAn and dx(n) gives, using (7.18.1) that en • en = 1,

dV = J en (Πi≠ndx'i) • (en dx'n) = J (Πidx'i) en • en = J (Πidx'i) (8.4.d.2)

which agrees with (8.4.a.1), verifying our assumption (8.4.d.1). Looking at dV = (dAn)i[dx(n)]i, dV is
seen to be a tensor combination of a vector density of weight -1 with a vector density of weight 0 (an
ordinary vector dx(n)), so according to (D.2.3) dV must be scalar density of weight -1. Rule (D.1.4) then
says,

dV' = J dV . (8.4.d.3)

Since dV = J (Πidx'i) from (8.4.a.1). one gets

dV' = J2 (Πidx'i) . (8.4.d.4)

Again, one can verify this last result from the x'-space covariant form of the equation:

dV' = dA'n • dx'(n) = { J2(Πi≠ndx'i) e'n } • {e'n dx'n) = J2(Πidx'i) e'n • e'n = J2(Πidx'i)
(8.4.c.5) (8.2.1) (8.4.d.5)
n n
since e' • e'n = e • en = 1.

150
 Chapter 8: Length Area and Volume

An alternate derivation of the dV transform rule dV' = J dV comes from just staring at (8.4.a.1),

dV = εabc..x (dx'e1)a(dx'e2)b..... (dx'eN)x (8.4.d.6)

which by the argument above is seen directly to transform as a tensor density of weight -1.
To complete the circle, we can verify for a second time the claim (8.4.d.1) that dV = dAn • dx(n) :

dAn • dx(n) = (dAn)i [dx(n)]i

= {(-1)n-1 εiabc..x (e1)a(e2)b.... (eN)x (Πk≠ndx'k) } (dx'nen)i // en missing .....

= εabc..i..x (e1)a(e2)b.. (en)κ.. (eN)x (Πkdx'i) // note indices on ε

= det(S) (Πkdx'i) // (5.12.8) where (eb)a = Sab from (7.18.1)

= dV . // from (8.4.a.1) (8.4.d.7)

To summarize the above, by examining the vector density nature of our various objects, we have been
able to determine exactly how edges, areas and the volume transform under transformation F:

edges dx'(n) = R dx(n) or [dx'(n)]i = Rij [dx(n)]j // ordinary vector

areas dA'n = J R dAn or (dA'n)i = J Rij (dAn)j // vector density W = -1
volume dV' = J dV // scalar density W = -1
(8.4.d.8)
(e) Covariant Magnitudes

The x'-space magnitudes here are the "covariant" ones which are associated with the Curvilinear View of
x'-space, as discussed above. Since dx(n) is a vector, it follows that

| dx'(n)|2 = dx'(n) • dx'(n) = dx(n) • dx(n) = | dx(n)|2 ⇒ | dx'(n)| = | dx(n)| . (8.4.e.1)

Since dAn is a vector density of weight -1, if follows that,

| dA'n |2 = dA'n • dA'n = J2 dAn • dAn = J2| dAn |2 ⇒ | dA'n | = |J| | dAn | (8.4.e.2)

where we note that dAn • dAn, being a combination of two weight -1 vector densities, is a scalar density
of weight -2 and thus transforms as shown above. For completeness, we can add from the above table,

V' = J dV ⇒ | dV' | = |J| | dV | . (8.4.e.3)

We now gather equations from above and take their absolute values on the right. Note from (7.18.1) that
|en| = h'n and |en| = g'nn .

151
 Chapter 8: Length Area and Volume

dx(n) = en dx'n (8.2.1) ⇒ |dx(n)| = |en| dx'n = h'n dx'n

dx'(n) = e'n dx'n (8.2.1) ⇒ |dx'(n)| = |e'n| dx'n = h'n dx'n

dAn = J (Πi≠ndx'i) en (8.4.a.1) ⇒ |dAn| = |J| |en| (Πi≠ndx'i) = |J| g'nn (Πi≠ndx'i)
dA'n = J2(Πi≠ndx'i) e'n (8.4.c.5) ⇒ |dA'n| = J2 |e'n| (Πi≠ndx'i) = J2 g'nn (Πi≠ndx'i)

dV = J (Πidx'i) (8.4.a.1) ⇒ |dV| = |J| (Πidx'i)

dV' = J2 (Πidx'i) (8.4.d.4) ⇒ |dV'| = J2 (Πidx'i) . (8.4.e.4)

The above can be written more compactly using the Cartesian-View coordinate variation groupings
shown in (8.2.3),

dx(n) = en dL'n ⇒ |dx(n)| = h'n dL'n dL'n ≡ dx'n

dx'(n) = e'n dL'n ⇒ |dx'(n)| = h'n dL'n `

dAn = J dA'n en ⇒ |dAn| = (|J| g'nn ) dA'n dA'n ≡ Πi≠ndx'i

dA'n = J2 dA'n e'n ⇒ |dA'n| = (J2 g'nn ) dA'n

dV = J dV' ⇒ |dV| = |J| dV' dV' ≡ Πidx'i

dV' = J2 dV' ⇒ |dV'| = J2 dV' . (8.4.e.5)

The covariant edge magnitude is of course unchanged by the transformation since it is a scalar, while the
area and volume magnitudes are magnified by |J| in going from x-space to x'-space. Since dV = dAn •
dx(n), it is clear that the area's transformation factor of |J| is passed onto the volume. Notice that dV' is
always positive regardless of the sign of J, and this is because x'-space is always a right-handed
coordinate system, as in Section 6.9. In contrast, dV can have either sign depending on the sign of J =
det(S), and so dV < 0 when the en form a left-handed coordinate system in x-space.

(f) Two Theorems : g'nn g' = cof(g'nn) and |(Πxi≠nei)| = cof(g'nn)

We now pause to prove two small theorems which will be used below.

Theorem 1: • g'nn g' = cof(g'nn) where g' = det(g'ij) (8.4.f.1)

• g'nn g' = g' / h'n2 (orthog only)

For orthogonal coordinates g'nn = 1/h'n2 as shown in (5.11.9), proving the second item The first equality
can be shown as follows. First, define these two regular matrices,

(g'up)ab ≡ g'ab (g'dn)ab ≡ g'ab .

Then since these metric tensors are inverses, use fact that A-1 = cof(AT)/detA to get

152
 Chapter 8: Length Area and Volume

g'up = (g'dn)-1 = cof(g'dnT)/det(g'dn) = cof(g'dn)/det(g'dn) // g'dn symmetric

⇒ (g'up)nn = cof[(g'dn)nn]/det(g'dn)
or
g'nn = cof[g'nn] / g' QED

Theorem 2: |(Πxi≠nei)| = cof(g'nn) (8.4.f.2)

The quantity on the left is this

|(Πxi≠nei)| ≡ | e1 x e2 ... x eN | where en is missing from cross product. (8.4.f.3)

We shall give three quick proofs of Theorem 2, the last being valid only for N=3.

• First, one form for dAn from (8.4.a.1) is this,

dAn = (-1)n-1 (Πxi≠nei) (Πi≠ndx'i) = (-1)n-1 (Πxi≠nei) dA'n

⇒ |dAn| = |(Πxi≠nei)| dA'n . (8.4.f.4)

Comparison of this last result with the third line of (8.4.e.5) shows that the following must be true.

|(Πxi≠nei)| = |J| g'nn = g'1/2 g'nn = cof(g'nn) . QED (8.4.f.5)

Note that since g=1 (Cartesian x-space), J2 = g' from (5.12.14), and the last step follows from Theorem 1.

• Here is a more direct proof: [ Πxi≠n (ei) is a vector in Cartesian x-space ]

| Πxi≠n (ei)|2 = Πxi≠n (ei) • Πxj≠n (ej) = [ Πxi≠n (ei)]k [ Πxj≠n (ej)]k (8.4.f.6)

= [ εkabc...x (e1)a(e2)b ...... (eN)x] [ εka'b'c'...x (e1)a'(e2)b' ...... (eN)x'] // en missing in both

= εkabc...x εka'b'c'...x {(e1)a(e2)b ...... (eN)x } (e1)a'(e2)b' ...... (eN)x' // en missing

= e1•e1 e2•e2 .... eN•eN + all signed permutations of the 2nd labels // en missing

= g'11g'22..... g'NN + all signed permutations of the 2nd indices, (7.18.1) // en missing

But this is last object is the determinant of the g'ij matrix with g'nn crossed out, which is to say, it is the
minor of g'nn. Since g'nn is a diagonal element, the minor and cofactor are the same. Thus, this last object
is in fact just cof(g'nn). QED.

153
 Chapter 8: Length Area and Volume

• A proof for N=3 uses normal vector algebra. Setting n = 1, for example, one needs to show that

| Πxi≠1 ei |2 = | e2 x e3 |2 = cof(g'11) .

To this end, use the vector identity

(A x B) • (A x B) = A2B2 – (A•B)2

and fact (7.18.1) that en • em = g'nm to show that,

| e2 x e3 |2 = (e2 x e3) • (e2 x e3 ) = |e2|2 |e3|2 - (e2•e3)2 = g'22 g'33 - (g'23)2 = cof(g'11).
(8.4.f.7)
and the cases n = 2 and 3 are similar.

(g) Cartesian-View Magnitude Ratios

In the Cartesian View of x'-space one can write the Cartesian x'-space magnitudes as

| dx'(n) |c = dL'n | dA'(n) |c = dA'n | dV' |c = dV' . (8.4.g.1)

Then from the three x-space magnitude equations in (8.4.e.5) one obtains the following three ratios of x-
space objects divided by their corresponding Cartesian-View x'-space objects:

| dx(n)|/ dL'n = h'n = [g'nn]1/2 = the scale factor for edge dx(n)

| dA(n)|/ dA'n = g'nn |J| = g'nn g'1/2 = [ g'nn g']1/2 = [cof(g'nn)]1/2 // Theorem 1 above

|dV| / dV' = |J| = g'1/2 . // g' ≡ det(g'ij) = J2 (8.4.g.2)

It is convenient to make the definition

dAn ≡ | dA(n)| (8.4.g.3)

and then the above area magnitude ratio relation may be written

dAn = cof(g'nn) dA'n (8.4.g.4)

and dA'n = Πi≠ndx'i is just a product of the appropriate curvilinear coordinate variations. Since g'nn is a
diagonal element of the g' matrix, one can reexpress the cofactor in the above equation in this manner:

cof(g'nn) = minor(g'nn) = det (g' with row n and column n crossed out) (8.4.g.5)

On the other hand, one can regard the number cof(g'nn) as one element of the cofactor matrix cof(g')
whose elements are given by [cof(g')]ij ≡ cof(g'ij). Then,

cof(g'nn) = [cof(g')]nn . (8.4.g.6)

154
 Chapter 8: Length Area and Volume

(h) Nested Cofactor Formulas and STS notation

Descending Leaf Hierarchy

The object dAn is the "area" of a face on a differential N-piped spanned by the full set of vectors ei but
not including en. This face, which is itself an (N-1)-piped, in turn has its own "areas" which are (N-2)-
pipeds, and so on, so there is a hierarchy of "areas" of dimensions N-1 all the way down. The area ratios
of corresponding areas under transformation F are determined by equations similar to (8.4.g.4). For this
purpose we define Cof to be a matrix and cof to be a number,

Cof(Mij) ≡ the submatrix of M obtained by crossing out row i and column j of M

cof(Mij) = (-1)i-j minor(Mij) = (-1)i-j det[Cof(Mij)] = the usual "cofactor" . (8.4.h.1)

Then

Cof(g'nn) = the submatrix of g' obtained by crossing out row n and column n of g'

cof(g'nn) = minor(g'nn) = det[Cof(g'nn)] . (8.4.h.2)

Then for example the mth face of face n of an N-piped has area ratio cof[Cof(g'nn)]mm . The object
inside the radical is the cofactor of the m,m element of the N-1 x N-1 matrix Cof(g'nn). If we refer to this
area as dAn,m we can go down the hierarchy in this manner

dAn = cof(g'nn) Πi≠ndx'i face n

dAn,m = cof[Cof(g'nn)]mm Πi≠n,mdx'i face m of face n

dAn,m,k = cof(Cof[Cof(g'nn)]mm)kk Πi≠n,m,kdx'i face k of face m of face n (8.4.h.3)

and so on. As a very simple example, for N = 3 the faces dAn,m are line segments dx'i and one has

⎛ g'11 g'12 ⎞
Cof(g'33) = g' g' ⇒ cof[Cof(g'33)]22 = g'11 = h'12 ⇒ cof[Cof(g'33)]22 = h'1
⎝ 21 22 ⎠
(8.4.h.4)
and dA3,2 = h'1 is in fact the edge length ratio given above in (8.4.g.2).

There is another way to write these area magnitudes if we assume that x-space is Cartesian, but it is
difficult to express in standard notation, so we momentarily revert to our developmental notation. In that
notation, if x-space is Cartesian we can write from (7.5.6) and (3.2.7) that ḡ' = STS,

155
 Chapter 8: Length Area and Volume

(8.4.h.5)

as was written earlier in (5.11.3). Since the covariant metric tensor ḡ'ij in developmental notation is
equal to g'ij in standard notation, we shall use the standard notation below in expressing these tensor
elements.

The area (magnitude) dAn of face n of a differential N-piped is, from (8.4.g.4) and (8.4.e.5),

dAn = cof(g'nn) Πi≠ndx'i . (8.4.h.6)

The object cof (g'nn) is the minor of the above bracketed matrix with row n and column n crossed out.
That reduced matrix can in fact be written

e1
⎛ e2 ⎞
Cof(g'nn) = S S )=⎜ ⎟
(n)T (n)
(e1, e2, ...eN) where en is missing from both vectors . (8.4.h.7)
..
⎝ eN ⎠
so

⎛ ee12 ⎞
cof(g'nn) = det[S(n)TS(n)] = det [ ⎜ ⎟ (e1, e2, ...eN) ] . (8.4.h.8)
..
⎝ eN ⎠
Therefore,

dAn = cof(g'nn) Πi≠ndx'i

= det[S(n)TS(n)] Πi≠ndx'i
where S(n) = (e1,e2....eN) with en missing . (8.4.h.9)

The same argument results in

dAn,m = cof[Cof(g'nn)]mm Πi≠n,mdx'i

= det[S(n,m)TS(n,m)] Πi≠n,mdx'i
where S(n,m) = (e1,e2....eN) with en and em missing . (8.4.h.10)

Continuing down one more level,

156
 Chapter 8: Length Area and Volume

dAn,m,k = cof(Cof[Cof(g'nn)]mm)kk Πi≠n,m,kdx'i

= det[S(n,m,k)TS(n,m,k)] Πi≠n,m,kdx'i
where S(n,m,k) = (e1,e2....eN) with en, em, ek all missing . (8.4.h.11)

In this descending hierarchy of n-piped "leaf" areas, we eliminate one tangent base vector ei at a time,
and the eliminated vectors' labels are used to label the "leaf" whose area is indicated, as in S(n,m).

Ascending Leaf Hierarchy

One can alternatively build up the hierarchy of areas from the bottom, for example using S[n.m] to
indicate the area of the 2-piped spanned by en and em in RN where the top level N-piped lives. One then
has

dA[n] = det(S[n]TS[n]) dx'n where S[n] = en . (8.4.h.12)

In this case S[n]T S[n] = en•en = h'n2 so dA[n] = h'ndx'n as expected. Next, for 2-pipeds,

dA[n.m] = det(S[n.m]TS[n.m]) dx'ndx'm where S[n.m] = (en, em) . (8.4.h.13)

In this case
⎛ en ⎞ (e , e )] = det ⎛ en•en en•em ⎞ = det ⎛ g'nn g'nm ⎞
det[S[n.m]TS[n.m]] = det [ ⎜ ⎟ ⎜ e •e e •e ⎟
⎝ em ⎠ n m ⎝ m n m m⎠ ⎝ g'mn g'mm ⎠

= g'nng'mm - (g'nm)2 = h'n2 h'm2 - (en•em)2 (8.4.h.14)

so
dA[n.m] = h'n2 h'm2 - (en•em)2 dx'ndx'm . (8.4.h.15)

Doing one more step, at the level of 3-pipeds in the hierarchy buildup one would then have

dA[n.m.k] = det[S[n.m.k]TS[n.m.k]] dx'ndx'mdx'k where S[n.m.k] = (en, em, ek) (8.4.h.16)

and now
⎛ en ⎞ ⎛ g'nn g'nm g'nk ⎞
det[S [n.m.k]T [n.m.k]
S ] = det [ m (en, em, ek)] = det⎜ g'mn g'mm g'mk
⎜ e ⎟ ⎟ . (8.4.h.17)
⎜ ⎟ ⎜ ⎟
⎝ ek ⎠ ⎝ g'kn g'km g'kk ⎠

Comment 1: Notice that the matrices S appearing above in the above STS structures are in general not
square matrices because the number of components N of the ei vectors in RN generally does not match the
number of ei vectors in the list which defines S. On the other hand, STS is always a square matrix and
therefore always has a determinant.

157
 Chapter 8: Length Area and Volume

Comment 2: We have used the term "leaf" to refer to any n-piped appearing in the top level N-piped, with
n then ranging from N down to 1 (which leaf is just point). In a more precise discussion of this subject,
these leaves are called "boundaries" and in addition to having "area", these boundaries have "orientation".
We have only discussed the orientation of the top level leaves by using the vector area dAn . Usually the
precise discussion is couched in the language of "differential forms", and by careful consideration of
boundaries and their orientations one is able to derive the generalized Stokes's theorem which reads

∫M dα = ∫∂M α (8.4.h.18)

where α is a differential n-form, dα is an (n+1)-form which is the "exterior derivative" of α, M is a smooth

type of surface called a manifold, and ∂M is the boundary of that manifold. This abstract theorem then
encompasses all the well-known integral theorems of analysis in any number of dimensions n. See
Sjamaar [2015] Sections 5.2 - 5.4 , Section 8.2, and Theorem 9.9 on page 117. The fact that det(STS) is
the area scaling factor appears as Theorem 8.4 on Sjamaar page 101 where it is called det(ATA) . See
also Appendix F (The Volume of an n-piped embedded in Rm) of Lucht Tensor Products.

(i) Transformation of arbitrary differential vectors, areas and volume

The above discussion is geared to the N-piped transform picture and reveals how dx(n), dA(n) and dV
transform where all these quantities are directly associated with the particular differential N-piped in x-
space spanned by the (endx'n) vectors.
But suppose dx is an arbitrary differential vector in x-space. Certainly dx' = R dx, so we know how
this dx transforms under F. But what about area and volume?
Based on the work in Appendix B as carried through into the N-piped transform discussion above, it
seems clear that an arbitrary differential area in x-space dA can be represented as follows : (g=1)

dA = (dx[1]) x (dx[2]) ... x (dx[N-1]) (8.4.i.1)

(dA)i = εiabc..x (dx[1])a(dx[2])b.... (dx[N-1])x , (8.4.i.2)

where the dx[i] are an arbitrary set of N-1 linearly independent differential vectors in x-space. For N=3
one would write dA = dx[1] x dx[2]. Since ε has weight -1 and all other objects are true vectors (weight
0), one again concludes that dA transforms as a tensor density of weight -1, so

dA' = J R dA or dA'i = J Rij dAj . (8.4.i.3)

If any of these dx[k] were a linear combination of the N-2 others, one could say dx[k] = Σjαkj dx[j] and
then the above (dA)i expression would give a sum of terms each of which vanishes by symmetry,
resulting in (dA)i = 0.
Finally, given the set of dx[i] linearly independent vectors shown above for i = 1,2..N-1, we can
certainly find one more such that all N are then linearly independent, so then we have a set of N arbitrary
differential vectors dx[i](arbitrary as long as they are linearly independent), and these will form a volume
in x-space,

158
 Chapter 8: Length Area and Volume

dV = det (dx[1], dx[2], ..... dx[N] ) = εabc..y (dx[1])a(dx[2])b.... (dx[N])y . (8.4.i.4)

By the argument just given, inspection shows that this dV transforms as a tensor density of weight -1 so
dV' = JdV.
These then are the transformation rules for arbitrary differential vectors, areas and volumes
transforming under F :

dx ' = R dx |dx'| = |dx|

dA' = J R dA |dA'| = |J| |dA|
dV' = JdV |dV'| = |J| |dV| . J = det(S) J2 = g' (8.4.i.5)

Review and covariant form of the dA and dV equations

To review, in Cartesian x-space we have these expressions for area and volume

dA = (dx[1]) x (dx[2]) ... x (dx[N-1])

dV = det [dx[1], dx[2], ... dx[N]] . (8.4.i.6)

Written out in terms of contravariant vector components (dx[i])j these expressions appear as,

(dA)i = εiabc..x (dx[1])a(dx[2])b.... (dx[N-1])x

dV = εabc..y (dx[1])a(dx[2])b.... (dx[N])y , (8.4.i.7)

where ε is the usual permutation tensor involved in cross products and determinants. As noted earlier,
these are both "valid tensor density equations" as in (7.15.9), so they can be written in x'-space as

(dA')i = ε'iabc..x (dx'[1])a(dx'[2])b.... (dx'[N-1])x

dV' = ε'abc..y (dx'[1])a(dx'[2])b.... (dx'[N])y . (8.4.i.8)

where dx'[i] = Rdx[i] and where ε' is the x'-space Levi-Civita tensor. From (D.5.13), each ε' can be
written as ε' = J2ε = (g'/g)ε = g'ε, so

(dA')i = g' εiabc..x (dx'[1])a(dx'[2])b.... (dx'[N-1])x

dV' = g' εabc..y (dx'[1])a(dx'[2])b.... (dx'[N])y . (8.4.i.9)

Since the permutation tensor now appears, these can be written in terms of cross products,

dA' = g' (dx'[1]) x (dx'[2]) ... x (dx'[N-1])

dV' = g' det [dx'[1], dx'[2], ... dx'[N]] . (8.4.i.10)

Using contravariant components [dx'[1]]i, this cross product and determinant are formed just as they are
in x-space. These two equations then give at least some feel for "the meaning of dA' and dV' in x'-space" .

Going back to x-space, we could have written the equations there using g = det(g'ij) = det(δij) = 1 :

159
 Chapter 8: Length Area and Volume

dA = g (dx[1]) x (dx[2]) ... x (dx[N-1])

dV = g det (dx[1], dx[2], ... dx[N]) . (8.4.i.11)

This then is a useful interpretation of the covariant form of these equations. Adding primes to everything
in the above two equations yields the previous two equations and only the permutation tensor ε is
involved in both sets of equations.

With this understanding, the above transformation rules for differential vectors, areas and volumes can be
extended from Picture B to the more general Picture A, where g is an arbitrary metric tensor,

How things look in developmental notation.

Recall that covariant tensor objects get overbars and all indices are down in the developmental notation
used in Chapters 1-6 of this document. Here then are some of the above equations expressed in this
notation:

d¯A = (dx[1]) x (dx[2]) ... x (dx[N-1])

dV = det [dx[1], dx[2], ... dx[N]] (8.4.i.12)

(d¯A)i = ε̄ iabc..x (dx[1])a(dx[2])b.... (dx[N-1])x

dV = ε̄ abc..y (dx[1])a(dx[2])b.... (dx[N])y (8.4.i.13)

(d¯A')i = ε̄ 'iabc..x (dx'[1])a(dx'[2])b.... (dx'[N-1])x

dV' = ε̄ 'abc..y (dx'[1])a(dx'[2])b.... (dx'[N])y (8.4.i.14)

dx ' = R d x |dx'| = |dx|

d¯A' = J ST d¯A |d¯A'| = |J| |d¯A|
dV' = JdV |dV'| = |J| |dV| J = det(S) J2 = g' (8.4.i.15)

Recall from (2.5.1) that a covariant vector transforms as V̄' = ST V̄, and a covariant vector density of
weight W will then transform as V̄' = J-W ST V̄ and this explains the middle line of the above three. The
contravariant differential area would transform as dA' = J R dA.

(j) Concatenation (Composition) of Transformations

Consider x" = F2(x') and x' = F1(x) so that x" = F2[F1(x)] ≡ F(x). At some point x, the linearization will
yield dx" = R2R1dx as the rule for vector transformation, so the R matrix associated with transformation

160
 Chapter 8: Length Area and Volume

F is R = R2R1, and then S = R-1 = R1-1R2-1 = S1S2. Each transformation will have an associated
Jacobian: J1 = det(S1) and J2 = det(S2). The concatenated transformation then has J = det(S) = det(S1S2)
= det(S1)det(S2) = J1J2. The implication is that when one concatenates (composes) two transformations in
this manner, the area and volume transformations shown above still apply, where J is taken to be the
product of the two underlying Jacobians.
For example, one could consider the mapping between two different skewed N-pipeds, each
representing a different curvilinear coordinate system, with our orthogonal N-piped as an intermediary
object,

(8.4.j.1)

In this case one has F = F2-1F1, so R = R2-1R1 = S2R1 and then S = S1R2 so J = J1/J2. This J then
would be used in the above area and volume transformation rules, for example, dAleft = J R dAright.
In the continuum mechanics flow application, g = 1 on both left and right as well as center, time t0 is
on the right, time t on the left, and the volume transformation is given by dVleft = J dVright where J is
associated with the combined F. This J then characterizes the volume change between an initial and final
flow particle where each is skewed in some arbitrary manner.

(k) Examples of area magnitude transformation for N = 2,3,4

In (8.4.g.4) it was shown that dAn = cof(g'nn) dA'n. Since this is a somewhat strange result, some
examples are in order. Recall that the dAn are the areas of the faces of the differential N-piped in x-space,
while the dA'n are the curvilinear coordinate variations one can visualize in the Cartesian-View picture
(8.2.2) shown above.

For N=2 the area magnitude transformation results are (for a general non-orthogonal x'-space system)

dA1 = g'22 dA'1 = h'2 dA'1 dA'1 = dx'1 = dL'1

dA2 = g'11 dA'2 = h'1 dA'2 dA'2 = dx'2 = dL'2 . (8.4.k.1)

These equations are simple because the area of a parallelogram "face" is the length of an edge and so
these equations just coincide with the length transformation results stated above . Remember that a face is
labeled by the index of the vector which does not span the face, so h2' appears in the face 1 equation.

For N=3 the area magnitude transformation results are

dA1 = g'22 g'33 - (g'23)2 dA'1 dA'1 = dx'2dx'3

dA2 = g'33 g'11 - (g'31)2 dA'2 dA'2 = dx'3dx'1
dA3 = g'11 g'22 - (g'12)2 dA'3 dA'3 = dx'1dx'2 . (8.4.k.2)

161
 Chapter 8: Length Area and Volume

For an orthogonal N=3 system the metric tensor g'ab is diagonal, and then the above simplifies to

dA1 = g'22 g'33 dA'1 = h'2 h'3 dA'1 dA'1 = dx'2dx'3

dA2 = g'33 g'11 dA'2 = h'3 h'1 dA'2 dA'2 = dx'3dx'1
dA3 = g'11 g'22 dA'3 = h'1 h'2 dA'3 dA'3 = dx'1dx'2 . (8.4.k.3)

For an N=4 orthogonal system (g'ij is diagonal with g'ii = h'i2),

dA1 = cof(g'11) dA'1 = h'2 h'3 h'4 dA'1 dA'1 = dx'2dx'3dx'4

dA2 = cof(g'22) dA'2 = h'1 h'3 h'4 dA'2 dA'2 = dx'3dx'4dx'1
dA3 = cof(g'33) dA'3 = h'1 h'2 h'4 dA'3 dA'3 = dx'4dx'1dx'2
dA4 = cof(g'44) dA'4 = h'1 h'2 h'3 dA'4 dA'4 = dx'1dx'2dx'3 . (8.4.k.4)

Example 2: Spherical Coordinates: area patches

Consider again dAn = cof(g'nn) dA'n. Since spherical coordinates are orthogonal, the orthogonal N=3
example above may be used. Example 2 of Chapter 5 showed in (5.13.15) that [ 1,2,3 = r,θ,φ ] ,

h'1 = h'r = 1 dA'1 = dx'2dx'3 = dθdφ

h'2 = h'θ = r dA'2 = dx'3dx'1 = drdφ
h'3 = h'φ = rsinθ dA'3 = dx'1dx'2 = drdθ (8.4.k.5)

Therefore

dA1 = dA1^
e1 ⇒ dAr = dAr ^
e r = dAr ^
r with dAr = h'2 h'3 dA'1 = r2sinθ dθdφ
dA2 = dA2^
e2 ⇒ e θ = dAθ ^
dAθ = dAθ ^ θ with dAθ = h'3 h'1 dA'2 = rsinθ drdφ
dA3 = dA3^
e3 ⇒ dAφ = dAφ ^ ^ with dAφ = h' h' dA'3 = rdrdθ
e φ = dAφ φ (8.4.k.6)
1 2

so that

dAr = r2sinθ dθdφ ^

r ρdφ rdθ ρ = rsinθ
θ ^
dA = rsinθ drdφ θ ρdφ dr
^
dAφ = rdrdθ φ rdθ dr (8.4.k.7)

where all three vectors are seen to have the correct dimensions length2. As an exercise in staring, the
reader is invited to verify these results from the picture below using the hints shown above on the right,

162
 Chapter 8: Length Area and Volume

(8.4.k.8)

8.5 Transformation of Differential Volume applied to Integration

As discussed in Section C.8, the integral ∫D dV h(x) is the same regardless of the way the dV elements
are chosen, as long as those elements exactly fill the integration region D.
In the discussion above, |dV| (call it dVa) refers to a positive differential volume element in x-space
which is typically not aligned with the axes and for a general transformation F is not in general
orthogonal, as shown on the right side of Fig (8.2.2). Moreover, the shape of the differential volume N-
piped varies over the region of integration. Nevertheless, this "rag-tag band" of differential volumes, as
mentioned below (C.8.3) for the 2D case, fills the integration region perfectly.
Alternatively one could consider |dV| (call it dVb) to be the usual dx1dx2.....dxN differential volume
elements in x-space, and of course this set of differential volume elements also fills the integration space
perfectly.
Thinking of these two different differential volumes as dVa and dVb, one can see from the definition
of the Riemann integral as the limit of a sum,

lim Σi dVa(xi) f(xi) = lim Σi dVb(xi) f(xi) (8.5.1)

that
∫D dVa h(x) = ∫D dVb h(x) . (8.5.2)

There would be little meaning to the statement dVa = dVb, since no one is claiming there is some
particular skewed N-piped of volume dVa which matches some axis-aligned N-piped of volume dVb .
Nevertheless, one could write dVa = dVb as a distributional symbolic equality where the meaning of that
symbolic equality is precisely the equivalence of the two integrals above for any domain D and for any
reasonable function h(x). [ Formally h(x) might have to be a "test function" φ(x). Certainly one would
require that both integrals converge. ]

It was shown in the second last line of (8.4.e.4) that the volume of the skewed N-piped on the right side of
Fig (8.2.2) was given by

dVa = |J(x')| dV' = |J(x')| ( Πi=1N dx'i) . |J(x')| = g'(x) // g = +1 (8.5.3)

163
 Chapter 8: Length Area and Volume

Combining this with the distributional symbolic equation dVa = dVb gives

dVa = dVb
or
|J(x')| ( Πi=1N dx'i) = ( Πi=1N dxi)
or
|J(x')| dV' = dVb . (8.5.4)

Now overriding our previous notation, we can make these new commonly used definitions

dV ≡ ( Πi=1N dxi)
dV' ≡ ( Πi=1N dx'i) (8.5.5)

and express the distributional result (8.5.4) as

|J(x')| dV' = dV . // |J(θ,ρ)| dρdθ = dxdy in (C.8.6) (8.5.6)

We refer to this distributional equality in the example of (C.8.6) as the "Jacobian Integration Rule". The
symbolic equation is a shorthand for this equation

∫D dV h(x) = ∫D' dV' |J(x')| h(x) (8.5.7)

where on the right h(x) = h(x(x')) and region D' is the same region as D expressed in terms of the x'
coordinates. Writing out the volume elements this says

∫D ( Πi=1N dxi) h(x) = ∫D' ( Πi=1N dx'i) |J(x')| h(x(x')) (8.5.8)

and finally, using J2 = g'/g = g' = det(ḡ') → det(g'ab) from (5.12.14),

∫D ( Πi=1N dxi) h(x) = ∫D' ( Πi=1N dx'i) [ det(g'ab) ] h(x(x')) . (8.5.9)

For example, when applied to polar and spherical coordinates, one gets

∫D dxdy h(x) = ∫D' drdθ [r] h(x(r,θ)) det(g'ab) = Πih'i = r (8.5.10)

∫D dxdydz h(x) = ∫D' drdθdφ [ r2sinθ ] h(x(r,θ,φ)) det(g'ab) = Πih'i = r2 sinθ (8.5.11)

In the first case h(x) = h(x,y) and h(x(r,θ)) = h(rcosθ,rsinθ).

In the second case h(x) = h(x,y,z) and h(x(r,θ,φ)) = h(rsinθcosφ,rsinθsinφ,rcosθ).

164
 Chapter 8: Length Area and Volume

8.6 Interpretations of the Jacobian

Using the facts of Section 5.12 (in Standard Notation) and of the above Sections, one can produce various
expressions and interpretations for the Jacobian J and its absolute value |J| :

J(x') ≡ det(Sij(x')) = det(∂xi/∂x'k) = 1/det(Rij(x(x')) = 1/ det(∂x'i/∂xk) (5.12.6)

|J(x')| = det(g'ab(x')) = g'(x') // g=1 (5.12.14)

|J(x')| = the volume of the N-piped in x-space spanned by the en(x), where x = F-1(x') (8.3.8)

|J(x')| = dVN-piped/dV' = ratio of differential x-space N-piped volume / ( Πi=1N dx'i) (8.4.e.5)

|J(x')| = dV/dV' = ( Πi=1N dxi)/ ( Πi=1N dx'i) // distributional Jacobian Integration Rule (8.5.6)

(8.6.1)

As discussed in Section 6.9, if the curvilinear coordinates are ordered so that the en form a right-handed
coordinate system, then det(S)>0, σ = sign(det(S)) = +1, and |J| = J. Some authors define their Jacobians
as the inverse of ours, see (5.12.3).

8.7 Volume integration of a tensor field under linear transformations

Recall the distributional statement of (8.5.6),

dV' = J-1 dV . (8.7.1)

In the language of Appendix D, this says that the volume element transforms from x-space to x'-space as a
scalar density of weight +1. Since by (D.1.6) g1/2 transforms as a scalar density of weight -1, according
to (D.2.3) the object g1/2dV then transforms as an ordinary scalar (weight 0) (see for example Weinberg
p 99 (4.4.6) ). Then if one were to define

Tijk... ≡ ∫D g1/2dV Aijk..(x) (8.7.2)

one might expect that, if Aijk..(x) transforms as a tensor field, then Tijk... might transform as tensor.
To investigate this conjecture, consider the above integral in x'-space,

T'ijk... ≡ ∫D' g'1/2dV' A'ijk..(x') = ∫D g1/2dV A'ijk..(x'(x)) // g1/2dV = scalar

= ∫D g1/2dV Rii'Rjj'..... Ai'j'k'...(x) .

165
 Chapter 8: Length Area and Volume

If the underlying transformation x' = F(x) is linear (see Section 2.8) then the Rab are independent of
position and we continue the above,

= Rii'Rjj'..... ∫D g1/2dV Ai'j'k'...(x)

= Rii'Rjj'..... Ti'j'k'... (8.7.3)

Therefore, for linear transformations x' = F(x), an integral of the form (8.7.2) of a tensor field is itself a
tensor of the same type.

Examples: Suppose g = 1 and x' = Rx where R is a global rotation, so we have a linear transformation.
For a rotation in developmental notation one has R-1 = RT. But (5.7.6) says g' = R g RT = R 1 R–1 = 1,
so g' = 1 as well. We may conclude that the volume integral of a tensor field of any type is a tensor of the
same type. Here are two simple examples:

Ji = ∫D dV xi

Jij = ∫D dV [ r2δij - xixj] . (8.7.4)

Under rotations, xi is a true vector, and r2δij - xixj is a true rank-2 tensor (traceless). It follows that Ji
and Jij are also tensors. Under rotations, mass density ρ transforms as a scalar, so the following objects
are tensors as well,

Ii = ∫D dV ρ(x) xi // vector

Iij = ∫D dV ρ(x) [ r2δij - xixj] . // rank-2 tensor, r2 = Σi(xi)2 (8.7.5)

Vector components Ii are the first moments of a mass distribution, while Iij is the usual inertia tensor.
Since Iij is real and symmetric, it can be diagonalized by a certain rotation R. We can think of this as
a transformation from x-space to x'-space where the resulting tensor I'ij is diagonal. In x'-space, the
diagonal elements of the tensor I'ij are then its eigenvalues (λi = I'ii), while the eigenvectors are the
axis-aligned e'i of Chapter 3. Back in x-space, the eigenvectors of the non-diagonal Iij are then en ≡ Se'n
where S = R-1. This can be verified as follows (developmental notation)

I' = R I R-1 ⇒ I'ij = Rii'Ii'j'(R-1)j'j = Rii'Rjj'Ii'j // contravariant tensor

I' e'n = λn e'n // I' is diagonal

so
I en = [ R-1I' R ] [Se'n] = R-1 I' (RS) e'n = S I' e'n = S λne'n = λn Se'n = λnen . (8.7.6)

166
 Chapter 9: Divergence

9. The Divergence in curvilinear coordinates

Note: Covariant derivations of all the curvilinear differential operator expressions appear in Chapter 15.
In Chapters 9 through 13, we provide more "physical" or "brute force" derivations which are, of necessity,
much less compact. That compactness is a testament to the power of the covariant derivative formalism,
which might be called "semicolon technology". The formalism is not used in Chapters 9 through 13.

9.1 Geometric Derivation of the Curvilinear Divergence Formula

In Cartesian coordinates div B = ∇ • B = ∂nBn, but expressed in curvilinear coordinates the right side has
a more complicated form.

We provide here a geometric derivation (in N dimensions) of the formula for the divergence of a
contravariant vector field expressed in curvilinear coordinates, which means x'-space coordinates with
Picture B.

(9.1.1)

This derivation is an exercise in using the transformation results obtained in Chapter 8 above, and in
understanding the meaning of the components of a vector, as discussed in Section C.5.

The divergence of a vector field can be computed in a Cartesian x-space by taking the limit of the flux
emerging from a closed volume divided by the size of the volume, in the limit that the volume shrinks
down around some point x. Being a scalar field, the divergence is a property of the vector field at some
point x and therefore cannot depend on the shape of the closed volume used for the calculation†. If the
shape of the volume is taken to be a standard-issue axis-aligned N-piped, the divergence obtained will be
expressed in terms of the Cartesian coordinates and in terms of the Cartesian components of the vector
field: [div B](x) = ∂nBn(x) where B = Bn^n. However, if the N-piped shape is the one below, evaluation of
this same [div B](x) produces an expression which involves only the curvilinear coordinates and the
curvilinear components of the vector field, as will now be demonstrated.

†
For example, the total amount of water per second flowing through a mathematical closed boundary
surrounding a point-source sprinkler head will not depend on the shape of that boundary. This fact is
part of the Divergence Theorem, see e.g. wiki.

We start by considering again our differential non-orthogonal N-piped sitting in x-space, which has edges
endx'n, faces dAn and volume dV, as discussed in Section 8.4 above:

167
 Chapter 9: Divergence

(9.1.2)

In order to avoid confusion with volume V or area A, we name the vector field B. As just noted, the
divergence of a vector field B is the total flux flowing out through the faces of the N-piped divided by the
volume of the N-piped, in the limit that all differentials go to 0. Thus one writes symbolically,

[div B](x) = (1/dV) ∫ dA•B = (1/dV) ∫ dA(x)•B(x) (9.1.3)

where the surface integral is over all the faces of the above x-space differential N-piped. Recall that as we
move around in space, the shape (and size) of the above N-piped changes, so the dA of a face changes,
hence dA(x).

Comment on the div B as a scalar. If B is a tensorial vector field, then we claim that div B is a tensorial
scalar field, and one can write

[div B]'(x') = [div B](x) . (9.1.4)

The operator object {(1/dV)∫dA(x)•} acts as a tensorial vector operator, so that the result of its action on
B in (9.1.3) is a tensorial scalar. In (8.4.c.2) and (8.4.d.3) it was shown that dA and dV are vector and
scalar densities of weight -1. This means that dV' = J dV and dA' = J RdA so the ratio dA/dV in our
operator is a tensorial vector.
The fact that div B is a tensorial scalar is much more obvious from the alternative divergence
derivation given in Section 15.3. There we find that div B = Bi;i which evaluated in Cartesian space
becomes div B = Bi,i = ∂iBi. But in this Chapter we are avoiding Appendix F covariant derivatives!

The task is now to compute the integral ∫dA(x)•B(x) over the N-piped in (9.1.2).

Appendix B shows that the N-piped faces come in parallel pairs, so we start by considering pair n. We
now restore the original Appendix B convention that An points outward from the N-piped far face
regardless of the sign of det(S). Then adding the differential distances (Πi≠n dx'i) to the first line of
(8.3.7) one gets,

168
 Chapter 9: Divergence

dAn(x) = |det(Sij(x'))| en(x) ( Πi≠n dx'i) = g'(x') en(x) ( Πi≠n dx'i) . (9.1.5)

From (5.12.6) and (5.12.14) with g = 1 one has |det(S)| = |J| = g'1/2, while en are the reciprocal base
vectors called En Chapter 6. Quantities dAn, S, en and g' are explicitly shown as functions of space, and
as usual x' = F(x). Again, the out-facing vector area for the far face of pair n is dAn, while the out-facing
vector area for the near face is - dAn .
From (9.1.5) the contribution to the above divergence integral from "far face n" is, approximately,

[dAn]•B(x) ≈ [ g'(x'far) en(xfar) ( Πi≠n dx'i) ]•B(xfar)

= g'(x'far) ( Πi≠n dx'i) en(xfar) • B(xfar) , (9.1.6)

where xfar is taken to be a point at the center of far face n. Recall from the 2nd last line of (7.13.10) that

B(x) = B'n(x')en where B'n(x') = en(x) • B(x) (9.1.7)

which says that, when B is expanded on the en, the coefficients B'n of the expansion are the contravariant
components of vector B transformed into B' in x'-space (the curvilinear coordinate space) by B' = RB.
Thus we write,

B'n(x'far) = en(xfar) • B(xfar) . (9.1.8)

Inserting this into (9.1.6) gives

dAn•B(xfar) ≈ g'(x'far) ( Πi≠n dx'i) B'n(x'far) . (9.1.9)

This far-face n contribution to the flux integral is now expressed entirely in terms of x'-space objects and
coordinates. A similar expression obtains for the near-face n, but the sign of dAn is reversed. Adding the
contributions of these two faces of pair n gives

∫two faces n dA•B(x) ={ g'(x'far) B'n(x'far) – g'(x'near) B'n(x'near) } ( Πi≠n dx'i) .

(9.1.10)
In x'-space, if x'near is a point at the center of the near face of face pair n, then

x'far = x'near + e'n dx'n where e 'n = axis-aligned basis vector in x'-space (9.1.11)

since these two points map into the near and far face-n centers in x-space. For any function f,

f(x'far) - f(x'near) ≈ (∂'n f(x'near)) dx'n // no implied sum on n (9.1.12)

where a change is made only in coordinate x'n by amount dx'n. Applying to f = J B'n yields

{ g'(x'far) B'n(x'far) – g'(x'near) B'n(x'near) } ≈ ∂ 'n [ g'(x'near) B'n(x'near)] dx'n . (9.1.13)

169
 Chapter 9: Divergence

In the limit dx'n → 0 one has xnear→ x. Then (9.1.10) with (9.1.13) gives

∫two faces n dA•B(x) = ∂'n [ g'(x') B'n(x')] dx'n ( Πi≠n dx'i)

= ∂ 'n [ g'(x') B'n(x')] ( Πi dx'i) . (9.1.14)

Now all N differentials are present in (Πidx'i). The total flux flowing out through all N pairs of faces of
the N-piped in x-space is this same result with an implicit sum on n, so

total flux = ∫ dA•B(x) = ∂ 'n [ g'(x') B'n(x')] ( Πi dx'i) = ∂ 'n [ g'(x') B'n(x')] dV ' (9.1.15)

where dV' = Πi dx'i is the volume of the differential N-piped in Cartesian-view x'-space as in (8.2.3). The
divergence of B from the defining symbolic expression is then

[div B](x) = ∫ dA•B(x) / dV = ∂ 'n [ g'(x') B'n(x')] (dV'/dV) (9.1.16)

where dV is the volume of the x-space N-piped shown in Fig (8.2.2). In (8.4.e.5) it is shown that

dV = (g')1/2 dV' ⇒ (dV'/dV) = 1/ g'(x') // (g')1/2 = J (9.1.17)

so that (9.1.16) becomes

[div B](x) = [1/ g'(x') ] ∂'n [ g'(x') B'n(x')] // all x'-space coordinates and objects (9.1.18)

[div B](x) = ∂nBn(x) . // all x-space coordinates and objects (9.1.19)

The added second line just shows [div B](x) expressed in terms of the Cartesian x-space coordinates and
objects, while the first line resulting from our derivation shows the same [div B](x) expressed in terms of
only x'-space objects and coordinates. The goal advertised above has been fulfilled.

If B is a tensorial vector, then as noted above div B is a tensorial scalar,

[div B](x) = [div B]'(x') (9.1.20)

and thus the left sides of both equations above could be replaced by [div B]'(x').

9.2 Various expressions for div B

It is shown above that

[div B](x) = [1/ g'(x') ] ∂'n [ g'(x') B'n(x')] . (9.2.1)

170
 Chapter 9: Divergence

To obtain div B written in terms of covariant components B'n, one sets B'n = g'nm B'm to get

[div B](x) = [1/ g'(x') ] ∂'n [ g'(x') g'nm(x') B'm(x')] . (9.2.2)

Recall that the B'm are the coefficients of B when expanded on the en, B = B'nen.

In practical work B is expanded on the unit vectors ^

e n ≡ en/ |en | = en/h'n so that

B = B'nen = B'n h'n^

e n = B'n^
en where B'n ≡ B'n h'n (9.2.3)

and then

[div B](x) = [1/ g'(x') ] ∂'n [ g'(x') B'n(x')/ h'n(x') ] . (9.2.4)

For example, spherical coordinate work might use ^e 1, ^

e 2, ^ r, ^
e3 = ^ ^. As noted earlier, the components
θ, φ
n
B (x) are not contravariant vector components since they don't quite transform properly:

B'n = RnmBm (B'n / h'n) = Rnm (Bn / 1) B'n = h'n (Rnm Bn) . (9.2.5)

Our Picture B results so far are these, assuming B is a tensorial vector,

General: B'n(x') = RnmBm(x) x = F-1(x') ≡ x(x') x' = F(x) ≡ x'(x)

[div B](x) = [1/ g'(x') ] ∂'n [ g'(x') B'n(x')] B = B'nen

[div B](x) = [1/ g'(x') ] ∂'n [ g'(x') B'n(x')/ h'n(x') ] B = B'n ^
en
[div B](x) = [1/ g'(x') ] ∂'n [ g'(x') g'nm(x') B'm(x')] B = B'nen
[div B](x) = ∂n Bn(x) // Bn = Cartesian components of B B = Bn ^
n (9.2.6)
[div B](x) = [div B]'(x')

For orthogonal curvilinear coordinates, one has

g'ij = h'i2 δi,j g'ij = h'i-2 δi,j det(g'ij) = Πih'i2 g' = (Πih'i) = h'1h'2....h'N (9.2.7)

so the above expressions can be written (the arguments x' of the h'n are now suppressed)

Orthogonal:

[div B](x) = [1/(Πih'i)] ∂'n [(Πih'i) B'n(x')] B = B'nen

[div B](x) = [1/(Πih'i)] ∂'n [(Πih'i) B'n(x')/ h'n ] B = B'n ^
en
[div B](x) = [1/(Πih'i)] ∂'n [(Πih'i) B'n(x')/h'n2] B = B'nen
[div B](x) = ∂n Bn(x) // Bn = Cartesian components of B B = Bn ^
n
[div B](x) = [div B]'(x') (9.2.8)

171
 Chapter 9: Divergence

Comment: Are the equations (9.2.6) valid if B is not a tensorial vector? For such a B one might try to
make it be tensorial "by definition" as discussed in Section 2.9. One would then go ahead and define
B'n(x') ≡ RnmBm(x) and claim success. If such a definition does not result in an inconsistency, then such a
B has been moved into the class of tensorial vectors. Example 1 in (2.9.3) shows have such an
inconsistency might arise, and it is interesting to see how that plays out here. Suppose F is non-linear so
that R and g' = RRT are functions of x' and are not constants. Take B(x) = x (the identity field) and try to
make it be contravariant by definition, x'n ≡ Rnmxm. The Cartesian divergence is then div B = ∂nBn(x) =
∂nxn = 3. But the first equation of (9.2.6) says

div B = [1/ g'(x') ] ∂'n [ g'(x') x'n] = ∂'n x'n + [1/ g'(x') ] x'n∂'n g'(x') = 3 + other stuff
(9.2.9)

and thus the two calculations for div B disagree. As noted earlier, x' ≡ Rx conflicts with x' = F(x) in the
case of non-linear F.
This comment can be applied to the tensorial character of the differential operators treated in later
Chapters.

9.3 Translation from Picture B to Picture M&S

(9.3.1)

Picture M&S reflects the notation used by Moon & Spencer. In order to avoid a symbol conflict with the
Cartesian tensor components, the Curvilinear (now u-space) components are displayed in italics.

The rules for translation are

• replace x' by u everywhere (9.3.2)

n n
• replace ∂'n by ∂n meaning ∂/∂u ( exception: on a "Cartesian" line ∂n means ∂/∂x )
• replace g' by g (both the scalar and the tensor) and hn' by hn
• put all primed tensor components (scalar, vector, etc) into unprimed italics (eg, B'n → Bn , f ' → f )

After this translation, all unprimed tensor components are functions of x, while all italicized tensor
components are functions of u.

Here then are the translations of the two blocks above: (implied summation everywhere)

172
 Chapter 9: Divergence

General: now Bn(u) = RnmBm(x) x = F-1(u) ≡ x(u) u = F(x) ≡ u(x)

[div B](x) = [1/ g ] ∂n [ g Bn] B = Bnen

[div B](x) = [1/ g ] ∂n [ g Bn/ hn ] B = Bn^
en // M&S 1.06
nm n
[div B](x) = [1/ g ] ∂n [ g g Bm] B = Bn e
[div B](x) = ∂nB n n
// B = Cartesian components of B B = Bn ^n= Bn ^
n
[div B](x) = [div B](u) // transformation (scalar) (9.3.3)

Orthogonal:

[div B](x) = [1/(Πihi)] ∂n [(Πihi) Bn] B = Bnen

[div B](x) = [1/(Πihi)] ∂n [(Πihi) Bn / hn] B = Bn ^
en
[div B](x) = [1/(Πihi)] ∂n [(Πihi) Bn / hn2] B = Bnen (9.3.4)

Notice that the scalar function [div B]'(x') → [div B](u) according to the fourth rule of (9.3.2) above,
and that the arguments of all hk(u) are suppressed.

As an example, for N=3 the second line of (9.3.4) becomes

[div B](x) = [1/(h1h2h3)] { ∂1[h2h3 B1(u) ] + cyclic } B = Bn ^

en (9.3.5)

where + cyclic means two other terms with 1,2,3 cyclically permuted.

With the replacements

B→E, Bn→ En hn → gnn

the 2nd equation of (9.3.3) agrees with Moon & Spencer p 2 (1.06).

Comment: We use the "script MT bold" font Bn for components of vectors expanded onto ^ e n. It had to be
something in upper case distinct from Bn and Bn. In practice one can replace Bn with a different symbol
and then Bn is just a formal notation appearing in formulas. For example, in spherical coordinates (1,2,3)
= (r,θ,φ) one can make the replacements B1, B2, B3 → Br, Bθ, Bφ and these then do not conflict with
Bx, By, Bz or Br, Bθ, Bφ . See Section 14.7 for another example. (9.3.6)

9.4 Comparison of various authors' notations

Different authors use different symbols for curvilinear coordinates. They usually use x-space as the
Cartesian space, and then something like u-space or ξ-space as the curvilinear space:

173
 Chapter 9: Divergence

Curvilinear coords Cartesian coords Curvilinear space gnn

Picture C xn x(0)n x-space hn
n
Picture B x' xn x'-space h'n
Moon & Spencer (M&S) p 2 un xn u-space gnn
Morse & Feshbach (M&F) p 115 ξn xn ξ-space hn
Margenau & Murphy p 192 qn xn q-space Qn

(9.4.1)

These authors don't use any special notation to distinguish Cartesian from curvilinear components, nor is
it always clear whether a component is a coefficient of a unit vector or not, so one must be careful. For
example, on page 115 Morse & Feshbach simply say

(9.4.2)
which compare to the 2nd line of (9.3.4) above (sum on n now displayed),

[div A](x) = [div A](u) = [1/(h1h2h3)] Σn ∂n [h1h2h3 An(u) / hn] A = An ^

en (9.4.3)

so one should identify the M&F An with An, the coefficient of ^

e n.

174
 Chapter 10: Gradient

10. The Gradient in curvilinear coordinates

10.1 Expressions for grad f

This Chapter is considered in the Picture B context,

(10.1.1)

The gradient of f is defined in Cartesian space by

[grad f]n ≡ Gn ≡ ∂nf(x) so G = grad f = ∂nf(x) ^

n . (10.1.2)

Comment: To follow our conventions, we should perhaps write [grad f] as [grad f] as in [∇f] with a
bolded del operator. We decided against this so that grad, div and curl are all written non-bolded. This
seems to be a common usage of other authors.

Assuming f is a tensorial scalar field under F, then the Gn = ∂nf(x) are covariant vector field components
under F. This is because,

[∂'nf'(x')] = (Rnm ∂m)f'(x') = (Rnm ∂m)f(x) = Rnm [∂mf(x)] , (10.1.3)

where recall from (2.10.1) that f'(x') = f(x) for a scalar field. [ In the technical language of Appendix F,
one can write from (F.9.1) that tensorial vector f;n = f,n = ∂nf. ]

Since ∂nf is a tensorial vector, the covariance of Section 7.15 gives us (10.1.2) expressed in x'-space,

[grad f]'n = G'n ≡ ∂'nf '(x') . (10.1.4)

According to the last line of (7.13.10) vector G can be expanded as

G = G'1e1 + G'2 e2 +... = ΣnG'n en where en • G = G 'n (10.1.5)

where the coefficients G'n are the covariant components of G' in x'-space, which is the transformed G.
One can thus write

G(x) = [grad f](x) = G'n en = ∂'nf '(x') en = en ∂'nf '(x') ≡ ∇'CLf '(x') ∇'CL ≡ en∂ 'n
G(x) = [grad f](x) = ∂nf(x) ^
n =^
n ∂nf(x) = ∇ f(x) ∇≡^ n ∂n (10.1.6)

where the second line shows the usual Cartesian form of the gradient. The first line shows how one could
define a "curvilinear gradient" operator ∇'CL ≡ en∂'n, but this does not seem particularly useful. To restate,

175
 Chapter 10: Gradient

[grad f](x) = ∂'nf '(x') en // Curvilinear

[grad f](x) = ∂nf(x) ^n . // Cartesian (10.1.7)

In the first line, the reciprocal vectors en exist in x-space, but the coefficients are expressed entirely in
terms of x'-space coordinates and objects. Since f(x) is a scalar field, f(x) = f '(x'), one could regard the
derivative appearing in the first line as

∂'nf '(x') = ∂'nf(x(x')) where x(x') = F-1(x') . (10.1.8)

The contravariant components of grad f are then easily obtained as

G'i(x') = [grad f]'i(x') = ∂'if '(x') = g'ij(x') ∂'jf '(x') G = grad f = G'i(x') ei (10.1.9)

where now the expansion G'i(x') ei is that of the 3rd line of (7.13.10).

If an expansion on unit vectors is desired, the right equation of (10.1.9) can be written, since ei = h'i^
ei ,

G = [grad f](x) = G'iei = (G'i h'i) ^

ei = G'i ^
ei where G'i ≡ h'i G'i (10.1.10)

so then (10.1.9) reads,

G'i(x')/h'i = [grad f]'i(x') = ∂'if '(x') = g'ij(x') ∂'jf '(x') G = grad f = G'i ^
ei . (10.1.11)

Gathering up these results one gets

G'i(x') = [grad f]'i(x') = ∂'if '(x') G = grad f = G'i(x') ei

G'i(x') = [grad f]'i(x') = ∂'if '(x') = g'ij(x') ∂'jf '(x') G = grad f = G'i(x') ei
G i(x') = h'i[grad f]'i(x') = h'i ∂'if '(x') = h'i g'ij(x') ∂'jf '(x') G = grad f = G'i(x') ^
ei
(10.1.12)
which can be rewritten

[grad f](x) = (∂'if '(x')) ei

[grad f](x) = (∂'if '(x')) ei = g'ij(x') (∂'jf '(x')) ei
[grad f](x) = h'i (∂'if '(x')) ^
e i = h'i g'ij(x') (∂'jf '(x')) ^
ei
[grad f](x) = (∂if(x)) ^
n = ∇f(x) // Cartesian
[grad f]'i(x') = Rij[grad f]j(x) // transformation f '(x') = f(x) (10.1.13)

Again, in each of the first three forms above, G = [grad f](x) is being expressed as a linear combination of
e vectors which are in x-space, but the coefficients are given entirely in terms of x'-space coordinates and
objects. One can compare the last line of (10.1.13) to the last line of (9.28) that [div B](x) = [div B]'(x'),
Since div B is a scalar, there is no mixture of components as in [grad f]'i(x') = Rij[grad f]j(x) .

176
 Chapter 10: Gradient

For the orthogonal case g'ij = (1/h'i2) δi,j from (5.11.9) so the block above becomes

[grad f](x) = (∂'if '(x')) ei

[grad f](x) = (∂'if '(x')) ei = (1/h'i2) (∂'if '(x')) ei
[grad f](x) = h'i (∂'if '(x')) ^
ei = (1/h'i) (∂'if '(x')) ^ei
[grad f](x) = (∂if(x)) ^
n // Cartesian
j
[grad f]'i(x') = Ri [grad f]j(x) // transformation f '(x') = f(x) (10.1.14)

This above equations can be converted from Picture B to Picture M&S using the same rules given in
(9.3.2), which we repeat below

(10.1.15)
• replace x' by u everywhere
• replace ∂'n by ∂n meaning ∂/∂un ( exception: on a "Cartesian" line ∂n means ∂/∂xn) (9.3.2)
• replace g' by g (both the scalar and the tensor) and hn' by hn
• put all primed tensor components (scalar, vector, etc) into unprimed italics (eg, B'n → Bn , f ' → f )

After this translation, all unprimed tensor components are functions of x, while all italicized tensor
components are functions of u.

The translated results are then (all implied sums)

[grad f](x) = (∂if ) ei

[grad f](x) = (∂if ) ei = gij (∂jf ) ei
[grad f](x) = hi(∂if ) ^
ei = hi gij (∂jf ) ^
ei
^
[grad f](x) = (∂if) n // Cartesian
j
[grad f ]i(u) = Ri [grad f]j(x) // transformation (vector) f (u) = f(x) = f(x(u)) (10.1.16)

Notice that [grad f]'i(x') → [grad f]i(u) according to the fourth rule, meaning G'i(x') → Gi(u) .

For orthogonal curvilinear coordinates,

[grad f](x) = (∂if ) ei

[grad f](x) = (∂if ) ei = (1/hi2) (∂if ) ei
[grad f](x) = hi(∂if ) ^
ei = (1/hi) (∂if ) ^ei // M&S 1.05 (10.1.17)

One can always make the replacement f (u) = f(x(u)) in any of the above equations (f scalar). And one
more time: the various e vectors are in x-space, but all the coefficients are expressed in curvilinear u-
space coordinates and components. With the replacements

177
 Chapter 10: Gradient

f→φ ^
e i→ a i hi → gii

the last equation of (10.1.17) agrees with Moon & Spencer p 2 (1.05).

10.2 Expressions for grad f • B

Sometimes one is interested in the following quantity (back to Picture B)

grad f • B (10.2.1)

where f is a tensorial scalar field and B is a tensorial vector field. In this case, since grad f is a tensorial
vector field, the quantity grad f • B is a tensorial scalar, and so (grad f)' • B' = (grad f) • B .

This quantity grad f • B can be written several ways depending on how B is expanded:

B = Σi B'i ei ⇒ grad f • B = (∂'nf ') en • Σi B'i ei = (∂'nf ') B'n

B = Σi B'i ei ⇒ grad f • B = (∂'nf ') en • Σi B'i ei = (∂'nf ') B'n . (10.2.2)

Here grad f comes from lines 2 and 1 of (10.1.14) and the B expansions from (7.13.10), and there is an
implied sum on n. The last line can be written, using B'i ≡ B'i h'i ,

B = Σi [B'i h'i] ^
e i ≡ Σi B'i ^
ei ⇒ grad f • B = (∂'nf ') B'n = (∂'nf ') (B'n/h'n) (10.2.3)

To summarize:

[grad f](x) • B(x) = (∂'nf '(x')) B'n(x') for B = Σi B'i ei

[grad f](x) • B(x) = (∂'nf '(x')) B'n(x') for B = Σi B'i ei
[grad f](x) • B(x) = (∂'nf '(x')) B'n(x')/h'n(x') for B = Σi B'i ^
ei
[grad f](x) • B(x) = (∂nf(x)) Bn(x) for B = Bn ^
n // Cartesian
[grad f](x) • B(x) = [grad f]'(x') • B'(x') // scalar (10.2.4)

Since grad f • B is a scalar, these results resemble the divergence results more than the gradient ones.
Everything on the right side of the first three equations involves only x'-space coordinates and
components.

The conversion from Picture B to Picture M&F is straightforward using rules (9.3.2),

[grad f](x) • B(x) = (∂nf ) Bn B = Σi Bi ei

[grad f](x) • B(x) = (∂nf ) Bn B = Σi Bi ei
[grad f](x) • B(x) = (∂nf ) Bn/hn B = Σi Bi ^
ei
[grad f](x) • B(x) = (∂nf) Bn // Cartesian B = Bn ^
n
[grad f](x) • B(x) = [grad f ](u) • B(u) // transformation (scalar) (10.2.5)

178
 Chapter 10: Gradient

where once again f (u) = f(x(u)).

Comment: According to the Cartesian line above, one can write

grad f • dx = ∂nf(x) dxn = df = f(x+dx) - f(x) . (10.2.6)

This equation df = [grad f](x) • dx is sometimes used as an alternate definition of the gradient. If dx is
selected to be in the direction of grad f, the dot product has its maximum value, and therefore the gradient
points in the direction of the maximum change of a scalar function f(x). For N=2, in the usual 3D plot of
real z = f(x,y) the gradient then points "uphill", and the negative of the gradient then points "downhill".

179
 Chapter 11: Laplacian

11. The Laplacian in curvilinear coordinates

We continue to use Picture B, see (11.9) below.

The Laplacian (also known as the Laplace-Beltrami operator) is defined by

lap f = div (grad f) = div G where G ≡ grad f . (11.1)

Since f is (by assumption) a tensorial scalar field, G = grad f is a tensorial vector, as shown in (10.1.3).
Then, as claimed in (9.1.4), div(grad f) = div G is a tensorial scalar, meaning [lap f](x) = [lap f]'(x').
[In Chapter 15 we shall find that lap f = f;i;i which in Cartesian space is f,i,i = ∂i∂if = ∂i2f. ]

In Cartesian coordinates one writes

lap f = ∇2f = ∇•∇f = Σn∂n2f (11.2)

but this form gets modified when lap f is expressed in curvilinear coordinates. The first line of (9.2.6)
showed that,

div G = [1/ g' ] ∂'m [ g' G' m] where G = G'nen g' = det(g'ab) . (11.3)

Meanwhile, the second line of (10.1.13) showed that,

grad f = G = [g'nm (∂'nf ') ] em = G' mem , G' m = g'nm (∂'nf ') , ∂'nf ' = ∂'n f(x(x')) = ∂'nf '(x') .
(11.4)
Therefore,

lap f = div (grad f) = div G = [1/ g' ] ∂'m [ g' G' m] = [1/ g' ] ∂'m [ g' g'nm (∂'nf ')] (11.5)

so the general results can be concisely stated:

[lap f](x) = [1/ g'(x') ] ∂'m [ g'(x') g'nm(x') (∂'nf '(x')) ] // implied sum on n and m
[lap f](x) = Σn ∂n2f(x) // Cartesian f '(x') = f(x) = f(x(x'))
[lap f](x) = [lap f]'(x') // scalar . (11.6)

For an orthogonal coordinate system we know from (5.11.9) that

g'nm = h'n2 δn,m g'nm = (1/h'n2) δn,m g' = Πi h'i (11.7)

so the first line of (11.6) simplifies to

[lap f](x) = [1/(Πih'i)] ∂'m [(Πih'i) (1/h'm2) (∂'mf ') ] . (11.8)

180
 Chapter 11: Laplacian

Converting from Picture B to Picture MS with rules (9.3.2) gives,

(11.9)
nm
[lap f](x) = [1/ g ] ∂m [ g g (∂nf ) ]
[lap f](x) = Σn ∂n2f(x) // Cartesian f (u) = f(x) = f(x(u))
[lap f](x) = [lap f ](u) // transformation (scalar) (11.10)

The first equation simplifies in the orthogonal case to

[lap f](x) = [1/ g ] ∂m [ g (1/hm2) (∂mf ) ] // orthogonal // M&S 1.09 (11.11)

or
[lap f](x) = [1/(Πihi)] ∂m [ (Πihi) (1/hm2) (∂mf ) ] . (11.12)

For N=3 this says,

[lap f](x) = 1/(h1h2h3) { ∂1 [ (h2h3/h1) ∂1f ] + cyclic } . (11.13)

With the replacements

f→φ hm2 → gmm

equation (11.11) agrees with Moon & Spencer p 3 (1.09).

181
 Chapter 12: Curl

12. The Curl in curvilinear coordinates

12.1 Definition of curl B

The vector curl is defined only in N=3 dimensions (but see Section 12.6 below). Picture B is used.

(12.1.1)

In Cartesian coordinates one writes

[curl B]i(x) = [∇ x B(x)]i = εijk∂jBk(x) , (12.1.2)

but when expressed in terms of curvilinear coordinates and components, the form is different.

Consider the x-space differential 3-piped shown on the right side of Fig (8.2.2),

(12.1.3)

This 3-piped has three pairs of parallel faces. Within each pair, the "near" face touches the point x at
which the tails of the spanning vectors meet, while the "far" face does not.

According to (8.4.a.1) we can write ( all dx'i > 0 )

dAn = J en (Πi≠ndx'i) . (12.1.4)

According to (8.3.6) the vector - dAn is the out-facing area vector for a near face. Therefore the in-facing
area vector for a near face is dAn. For example, the area vector dA3 of the bottom face in (12.1.3) points
up, because the bottom face is a near face and in-facing for it means up.

182
 Chapter 12: Curl

Consider now the line integral of a vector field B(x) around the boundary of near face n, where the
circulation sense of the integral is determined from the right-hand-rule by the direction of dAn. Denote
this line integral by

(∫
{ B•dx)n . (12.1.5)

Sometimes this line integral is referred to as "the circulation" or "the rotation" of B around near face n
(and rot B is another notation used for curl B). This terminology is suggestive of a fluid vortex where the
fluid velocity v has a large curl component perpendicular to the vortex.

In x-space the quantity C(x) ≡ curl B(x) is a vector field defined in the following manner in the limit that
all the differentials dx'n → 0 :

C • dAn = ( ∫
{ B•dx)n C ≡ curl B , (12.1.6)

Since dAn is given in (12.1.4) in terms of en, C should be expanded on the ek. Since C is a vector density
of weight -1, (D.2.9) shows that the proper expansion of C onto the ek is given by

C = J-1 Σk=13 C'k ek (12.1.7)

so that, using (12.1.4),

C • dAn = J-1 (Σk C'k ek) • ( J en ( Πi≠n dx'i) ) = C'n ( Πi≠n dx'i) . (12.1.8)

Combining this with (12.1.6) gives,

C'n(x') ( Πi≠n dx'i) = ( ∫

{ B•dx)n . (12.1.9)

Since (Πi≠ndx'i) = dAn is the area of face n of our differential N-piped, (12.1.9) says that the nth
component of the curl in x'space C'n equals the line integral of B around the boundary of face n, divided
by the area of face n, in the limit that all differentials go to zero.

Our task is now to compute this line integral and thereby come up with an expression for C'n(x'), the
components of C = curl B when B is expanded onto the ek in x-space.

12.2 Computation of the line integral

This shall be done for the bottom face (n=3) of the x-space differential 3-piped. As noted below (12.1.4),
the area vector dA3 points up in Fig (12.1.3). Our circulation integral by the right hand rule will be
counterclockwise around this bottom face boundary as seen in the figure, replicated here,

183
 Chapter 12: Curl

(12.2.1)

In this picture, it is intended that the upper right face is farthest from the viewer, so it is the "back" face of
the 3-piped. It's parallel partner face touching the point x is the "front" face. We could call these faces
near and far, but front and back seem better here. Then the bottom face about which we are integrating
has edges front, right, back, and left.

So, grouping the contributions in pairs, one sees that.

(∫
{ B•ds)3 ≈ [B(xfront) - B(xback)] • (e1 dx'1) + [B(xright) - B(xleft)] • (e2 dx'2) (12.2.2)

where B(xfront) refers to the value of B at the center of the front edge of the parallelogram which is the
bottom face, and similarly for the other three edges. In the limit that the dx'n → 0, this simple
approximation of the line integral is "good enough" to produce the desired results.
Motivated by (7.18.1) that ei • ej = δij, expand vector B as in the last line of (7.13.10),

B = B'jej where B'j(x') = B(x) • ej (12.2.3)

where the B'j are the covariant components of B in x'-space. Inserting this expansion four times into
(12.2.2) gives,

(∫
{ B•dx)3 = [B'1(x'front) - B'1(x'back)] dx'1 + [B'2(x'right) - B'2(x'left)] dx'2 (12.2.4)

where x'front = F(xfront) and similarly for the other three points. In x-space the figure shows,

dxBF ≡ xback - xfront = e2dx'2

dxRL ≡ xright - xleft = e1dx'1 . (12.2.5)

Applying matrix R gives the corresponding x'-space equations (recall dx' = Rdx and e'n = Ren),

dx'BF ≡ x'back - x'front = e'2dx'2 e'2 = (0,1,0...)

dx'RL ≡ x'right - x'left = e'1dx'1 e'1 = (1,0,0...) . (12.2.6)

184
 Chapter 12: Curl

Using the fact that

f(x'+dx') ≈ f(x') + (∂f/∂x'n) dx'n

for a variation dx' = e'ndx'n (with no sum on n), one writes,

B'1(x'back) ≈ B'1(x'front) + (∂B'1/∂x'2) dx'2 where dx' = e'2dx'2 n=2

B'2(x'right) ≈ B'2 (x'left) + (∂B'2/∂x'1) dx'1 where dx' = e'1dx'1 n=1 . (12.2.7)

The circulation integral (12.2.4) is then

(∫
{ B•dx)3 ≈ [– (∂B'1/∂x'2) dx'2] dx'1 + [(∂B'2/∂x'1) dx'1] dx'2

= [– (∂B'1/∂x'2) + (∂B'2/∂x'1)] dx'1 dx'2 = [– ∂'2B'1 + ∂'1B'2] dx'1 dx'2

= [∂'1B'2 – ∂'2B'1] dx'1 dx'2

= ε3ab ∂'aB'b ( Πi≠3 dx'i) . (12.2.8)

Repeating this calculation for faces 1 and 2 produces cyclic results, and all three face line integrals can be
summarized as (where equality holds in the limit dx'i → 0)

(∫
{ B•dx)n = εnab ∂'aB'b ( Πi≠n dx'i) . (12.2.9)

Appendix D discusses the tensor ε known as the Levi-Civita ε tensor. In Cartesian space, the up and down
position of the indices does not matter, as for any tensor. In non-Cartesian space up and down does
matter, as with any tensor. The only fact needed here is (D.4.8) that ε'abc.. = εabc.. where ε' is the
tensor in x'-space. In Cartesian space one can regard εabc.. = εabc.. as a bookkeeping permutation tensor
with the properties stated below (7.7.4) Installing then εnab = ε'nab into (12.2.9) yields,

(∫
{ B•dx)n = ε'nab ∂'aB'b ( Πi≠n dx'i) (12.2.10)

and this integral is then given entirely in terms of x'-space coordinates and objects.

12.3 Solving for the curl

Recall the expression (12.1.9) involving the curl component C'n in x'-space,

C'n(x') ( Πi≠n dx'i) = ( ∫

{ B•dx)n . (12.1.9)

Insert (12.2.10) to get

185
 Chapter 12: Curl

C'n(x') ( Πi≠n dx'i) = ε'nab ∂'aB'b ( Πi≠n dx'i) . (12.3.1)

The differentials cancel out, so then take dx'i→ 0 and thus shrink the 3-piped down to the point of
interest x = F-1(x') so that

C'n = ε'nab ∂'aB'b . (12.3.2)

Then from (12.1.7) we find, since J = g' ,

curl B = C = J-1 C'n en // implied sum on n

= [(1/ g' ) ε'nab ∂'aB'b ] en . (12.3.3)

The comparison between the curvilinear- and Cartesian-expressed curls is this:

[curl B](x) = [(1/ g' ) ε'nab ∂'aB'b(x') ] en = (1/ g' ) { [∂'1B'2 - ∂'2B'1] e3 + cyclic }

[curl B](x) = εnab ∂aBb(x) ^

n = { [ ∂1B2 - ∂2B1 ] ^
3 + cyclic } . (12.3.4)

Comment 1. The equation curl B = C = [(1/ g' ) ε'nab ∂'aB'b ] en obtained above assumed that B was a
vector and the expansion B = B'nen was used. If B were a vector density of weight -1, the expansion
would be B = J-1B'nen and the result would be curl B = C = [(1/ g' ) ε'nab ∂'a(J-1B'b) ]. This situation
will arise in consideration of the vector Laplacian in Chapter 13. Once again, J = g' . (12.3.5)

Comment 2. Our "big result" of this Section is really (12.3.2) that C'n = ε'nab ∂'aB'b. We could have
obtained this same result starting way back with (12.1.2) that Cn = εabc∂aBb . The incantation is that this
latter equation, being a tensor density equation, is covariant in the sense of (7.15.9) and therefore in x'-
space it has the same form with everything primed. But doing it the hard way provides a good physical
intuition for the curl of a vector field. (12.3.6)

12.4 Various forms of the curl

In (12.3.2) and (12.3.3) we found that,

C'n = ε'nab ∂'aB'b C = J-1C'n en B = B'n en

curl B = [(1/ g' ) ε'nab ∂'aB'b ] en . (12.4.1)

If it is desired to have contravariant components of B, one gets

C'n = ε'nab ∂'a(g'bcB'c ) C = J-1C'n en B = B'n en

curl B = [(1/ g' ) ε'nab ∂'a(g'bcB'c )] en . (12.4.2)

For practical applications, one usually wants both vectors expanded on the ^
e n unit vectors in this way

186
 Chapter 12: Curl

C = J-1C'n en = (J-1C'n h'n) ^

e n ≡ C 'n ^
en C 'n = h'n J-1C'n
B = B' en = (B' h'n) ^
n n
e n ≡ B' ^n
en B'n = h'nB'n . (12.4.3)

Using then C 'n = (J-1h'n)C'n and C = (J-1h'n)C'n ^

e n and J = g' one finds,

C'n = [(1/ g' ) h'n ε'nab ∂'a(g'bc B'c/h'c )] C = C 'n ^

en B = B 'n ^
en
curl B = [(1/ g' ) h'n ε'nab ∂'a(g'bc B'c/h'c )] ^
en curl B = C . (12.4.4)

To summarize: B'c = RcdBd g' = g'(x') etc.

[curl B](x) = ε'nab [(1/ g' ) ∂'aB'b ] en B = B'nen (12.4.1)

nab c n
[curl B](x) = ε' [(1/ g' ) ∂'a(g'bcB' )] en B = B' en (12.4.2)
[curl B](x) = ε' [(1/ g' ) h'n ∂'a(g'bc B' /h'c )] ^
nab c
en B = B' ^en n
(12.4.4)
[curl B](x) = εnab
∂aBb(x) ^n // Cartesian n^
B=B n (12.4.5)

Converting from Picture B to Picture MS with rules (9.3.2),

(12.4.6)
one gets :

[curl B](x) = εnab [(1/ g ) ∂aBb ] en B = Bnen

[curl B](x) = εnab [(1/ g ) ∂a(gbcBc )] en B = Bnen
[curl B](x) = εnab[(1/ g ) hn ∂a(gbc Bc/hc )] ^
en B = Bn^
en
[curl B](x) = εnab
∂aBb(x) ^n // Cartesian B = Bn ^
n (12.4.7)

Warning: The object εnab in the first three equations is now in u-space which is non-Cartesian, so up and
down index positions do matter, but when indices are all up, it continues to be the normal permutation
tensor.

Determinant Notation

Consider the following determinant,

⎪ X1 X2 X3 ⎪
det(M) = ε abc
XaYbZc = ⎪ Y1 Y2 Y3 ⎪ .
⎪ ⎪
⎪ Z1 Z2 Z3 ⎪

187
 Chapter 12: Curl

If one regards Xn as a vector Xn with components (Xn)i one could write the above line as three
determinant equations,

⎪ (X1)i (X2)i (X3)i ⎪

Qi = det(Mi) = ε abc
(Xa)iYbZc = ⎪ Y1 Y2 Y3 ⎪ , i = 1,2,3 .
⎪ ⎪
⎪ Z1 Z2 Z3 ⎪

One could then combine these into a single vector equation,

⎪ X1 X2 X3 ⎪
Q = ⎪ Y1 Y2 Y3 ⎪ = εabc XaYbZc .
⎪ ⎪
⎪ Z1 Z2 Z3 ⎪

As an example, suppose Xn = en , Yn = ∂n (an operator), and Zn = Bc . Then one has

⎪ e1 e2 e3 ⎪
Q = ⎪ ∂1 ∂2 ∂3 ⎪ = εabc X Y Z = εnab X Y Z = εnab e (∂ B ) .
⎪ ⎪ a b c n a b n a b
⎪ B1 B2 B3 ⎪

But the expression on the right is recognized as g [curl B](x) from the first line of (12.4.7). It is in this
notation that we now rewrite (12.4.7) in the traditional determinant notation:

⎪ e1 e2 e3 ⎪
[curl B](x) = (1/ g ) ⎪ ∂1 ∂2 ∂3 ⎪ B = Bn en (12.4.8)
⎪ ⎪
⎪ B1 B2 B3 ⎪

⎪ e1 e2 e3 ⎪
[curl B](x) = (1/ g ) ⎪ ∂1 ∂2 ∂3
⎪ B = Bn en (12.4.9)
⎪ c c c ⎪
⎪ g1cB g2cB g3cB ⎪

⎪ h1 ^e1 h2 ^e2 h3 ^e3 ⎪

[curl B](x) = (1/ g ) ⎪ ∂1 ∂2 ∂3 ⎪ B = Bn ^
en (12.4.10)
⎪ c c c ⎪
⎪ (g1c/hc) B (g2c/hc) B (g3c/hc) B ⎪

⎪ ^
1 ^ 2 ^3 ⎪
[curl B](x) = ⎪ ∂1 ∂2 ∂3 ⎪ // here ∂n= ∂/∂xn and Bi = Bi(x) B = Bn ^
n (12.4.11)
⎪ ⎪
⎪ B1 B2 B3 ⎪

12.5 The curl in orthogonal coordinate systems

For such systems gij = δi,jhi2 and det(gab) = h12h22h32 so g = h1h2h3 . It is then a simple matter to
convert all the above forms and the results are:

188
 Chapter 12: Curl

Picture B: B'c(x') = RcdBd(x) hi' = hi'(x') etc.

[curl B](x) = ε'nab [(1/(h1'h2'h3') ∂'aB'b ] en B = B'nen

[curl B](x) = ε'nab [(1/(h1'h2'h3')) ∂'a(h'b2B'b )] en B = B'nen
[curl B](x) = ε'nab [(1/(h1'h2'h3')) h'n ∂'a(h'b B'b) )] ^
en B = B 'n ^
en
[curl B](x) = εnab ∂aBb(x) ^ n // Cartesian B = Bn ^
n (12.5.1)

Picture M&S (12.4.6) : Bc(u) = RcdBd(x) hi = hi(u) etc.

[curl B](x) = εnab [(h1h2h3)-1 ∂aBb ] en B = Bnen

[curl B](x) = εnab [(h1h2h3)-1 ∂a(hb2Bb )] en B = Bnen
[curl B](x) = εnab[(h1h2h3)-1 hn ∂a(hb Bb)] ^ en B = Bn^
en (12.5.2)

⎪ e1 e2 e3 ⎪
[curl B](x) = (h1h2h3) -1 ⎪ ∂1 ∂2 ∂3 ⎪ B = Bn en (12.5.3)
⎪ ⎪
⎪ B1 B2 B3 ⎪

⎪ e1 e2 e3 ⎪
[curl B](x) = (h1h2h3) -1 ⎪ ∂1 ∂2 ∂3 ⎪ B = Bn en (12.5.4)
⎪ 2 1 2 2 2 3 ⎪
⎪ h 1 B h2 B h3 B ⎪

⎪ h1 ^e 1 h2 ^e 2 h3 ^e 3 ⎪
-1⎪
[curl B](x) = (h1h2h3) ∂1 ∂2 ∂3 ⎪ B = Bn ^
en // M&S 1.07a
⎪ 1 2 3 ⎪
⎪ h1 B h2 B h3 B ⎪
(12.5.5)
With the replacements

B→E Bn→ En ^
e n→ a n hi → gii (h1h2h3)-1 → (1/ g )

equation (12.5.5) agrees with Moon & Spencer p 3 (1.07a).

12.6 The curl in N > 3 dimensions

Looking at the basic form (12.4.7) of the curl,

[curl B]n(x) = εnab ∂aBb(x) // Cartesian (12.6.1)

it is hard to imagine a generalization to N>3 dimensions where the curl is still a vector. The only vectors
available for construction purposes are ∂n and Bn. For N=4 one might try out various generalizing forms

[curl B]n(x) = (1/ g ) εnabc ∂a(∂b Bc) = (1/ g ) εnabc ∂a∂b Bc ?

[curl B]n(x) = (1/ g ) εnabc ∂a (BbBc)) ?

189
 Chapter 12: Curl

but these two forms vanish because antisymmetric ε is contracted against something symmetric. Thus the
idea of using multiple cross products as used in Appendix A does not prove helpful.

At this point we call upon the Appendix F notion of covariant derivatives and related semicolons.

The rank-2 tensor Bb;a – Ba;b = ∂aBb – ∂bBa discussed in (D.8.5) provides the logical extension of the
curl to N > 3 dimensions. For N=3 it happens that the object can be associated with a vector curl,

[curl B]n = εnab [Bb;a – Ba;b ]/2 = εnab Bb;a = εnab [∂aBb – ∂bBa ]/2 = εnab∂aBb . (12.6.2)

In relativity where N=4, there is no vector curl, and one sees Bb;a – Ba;b referred to as the covariant curl,
and ∂aBb – ∂bBa as the ordinary curl (Weinberg p 106 (4.7.2)).
Writing the N-dimensional contravariant curl components in this manner in Cartesian x-space,

[curl B]ij = (Bj;i – Bi;j) (12.6.3)

one can then ask how this generalized curl would be expressed in terms of x'-space coordinates and
objects. (This curl is a regular rank-2 tensor with weight 0, the vector curl has weight -1 ).
The general issue of expanding tensors is addressed in Appendix E where this general result is
obtained in (E.2.11),

A = Σijk... A'ijk... (ei⊗ej⊗ek...) . A'ijk... = contravariant components of A in x'-space

(12.6.4)
Applying this to A = [curl B] one gets

[curl B] = Σij[curl B]'ijei⊗ej = Σij(B'j;i – B'i;j)ei⊗ej . (12.6.5)

The Cartesian components of curl B in x-space can then be expressed in terms of x'-space components
and coordinates,

[curl B]ab(x) = Σij [curl B]'ij (ei)a(ej)b

= Σij(B'j;i – B'i;j)(ei)a(ej)b (12.6.6)

where the tangent base vectors en exist as usual in x-space, and B'i = B'i(x') where x' = F(x). Using
(F.9.5) one finds that

(B'a;b – B'b;a) = ( ∂'aB'b - ∂'bB'a ) + ( g'acΓ'bcn - g'bcΓ'acn)B'n (12.6.7)

where Γ is the affine connection. As in previous Sections we could write this in various ways involving
components B'n, B'n, and B 'n. Since (12.6.7) is covariant, in x-space it takes the same form with no
primes

(Ba;b – Bb;a) = ( ∂aBb - ∂bBa ) + ( gacΓbcn - gbcΓacn)Bn (12.6.8)

= ( ∂aBb - ∂bBa ) .

190
 Chapter 12: Curl

The second line is due to having Γ = 0 by (F.4.16) for x-space being Cartesian with g = 1. This is also true
for a quasi-Cartesian x-space having g = G, as discussed in (1.10), because the x-space basis vectors are
then constants and Γ involves their spatial derivatives.

191
 Chapter 13: Vector Laplacian

13. The Vector Laplacian in curvilinear coordinates

13.1 Derivation of the Vector Laplacian in general curvilinear coordinates

The vector Laplacian is defined in terms of the vector curl which is only defined for N=3 dimensions.
The context is Picture B,

(13.1.1)

The definition of the vector Laplacian of a vector field B(x) is

∇2B ≡ grad(div B) – curl (curl B) . (13.1.2)

If B is a vector, div B is a scalar, grad(div B) is a vector, so we expect ∇2B to be a vector. An annoying

observation is that, according to Section D.8, curl B is a vector density of weight -1 and therefore
curl(curlB) a vector density of weight -2, and so apples are being subtracted from oranges to make ∇2B.
This tensor conundrum is resolved in Section 15.7 so we ignore it for now and proceed undaunted.

In Cartesian coordinates, one finds that, as proven below (15.7.3),

[∇2B]i = ∇2(Bi) ≡ Σn ∂n2Bi , (13.1.3)

but expressed in general curvilinear coordinates the form gets modified.

To avoid confusion, some authors use different symbols for the vector Laplacian operator. For example,
Moon and Spencer use @ in place of ∇2 and we will honor these authors by using that symbol here, so

@B ≡ grad(div B) – curl (curl B) . (13.1.4)

In order to make use of the results of earlier Sections, define

G ≡ grad(f) where f = div B (13.1.5)

V ≡ curl C where C ≡ curl B (13.1.6)

so that

@B = G – V . (13.1.7)

The first line of (10.1.13) gives this expression for G,

192
 Chapter 13: Vector Laplacian

G = grad(f) = (∂'kf ') ek . (13.1.8)

Now replace f 'using the first line of (9.2.6)

f ' = f '(x') = f(x) = div B = [1/ g' ] ∂'i [ g' B'i] (13.1.9)

so that

G = ∂'k{ (1/ g' ) ∂'i ( g' B'i)} ek

= ∂'n{ (1/ g' ) ∂'i ( g' B'i)} en . (13.1.10)

The second term V is little more complicated. First, from the first line of (12.4.5) write,

C = curl B = ε'nab [(1/ g' ) ∂'a{ B'b} ] en = J-1C'n en = (1/ g' )C'n en // J = g'

V = curl C = ε'ncd [(1/ g' ) ∂'c{ J-1C'd} ] en = J-2V'n en = (1/g') V'n en (13.1.11)

where recall from (D.4.8) that ε'abc... = εabc.. = the usual permutation tensor, but written up and
primed so as to be in covariant form. The factors of J appearing in the expansions on the far right are due
to the fact that C is a vector density of weight - 1 and V a vector density of weight -2, see (D.2.9).

From (13.1.11) we extract these facts,

C'n = ε'nab ∂'a{ B'b} (13.1.12)

n ncd
V' = g' ε' ∂'c{ (1/ g' )C'd} . (13.1.13)

Now lower the index on C'n in (13.1.12) to get

C'd = g'deC'e = g'de ε'eab (∂'aB'b) , (13.1.14)

and insert this into (13.1.13) to get

V'n = g' ε'ncd ∂'c{ (1/ g' ) g'de ε'eab (∂'aB'b) }

= g' ε'ncd ε'eab ∂'c { (1/ g' ) g'de(∂'aB'b) } . (13.1.15)

Then from (13.1.11) we have this expression for V,

V = curl C = (1/g') V'n en = [ (1/ g' )ε'ncd ε'eab ∂'c { (1/ g' ) g'de(∂'aB'b) }] en . (13.1.16)

We can then summarize our results:

193
 Chapter 13: Vector Laplacian

@B = G – V (13.1.7)
G = ∂'n{ (1/ g' ) ∂'i ( g' B'i)} en (13.1.10) B = B'nen
V = [ (1/ g' ) ε'ncd ε'eab ∂'c { (1/ g' ) g'de(∂'aB'b) }] en (13.1.16) B = B'nen
= [ (1/ g' ) ε'ncd ε'eab ∂'c { (1/ g' ) g'de(∂'a[g'bfB'f]) }] en . B = B'nen (13.1.17)

Often one wants vectors expanded onto the unit vectors ^

en

B = B'n en = B'n h'n ^

e n = B 'n^
en ⇒ B'n = B 'n/h'n (13.1.18)

and the above set of equations becomes

@B = G – V
G = h'n ∂'n{ (1/ g' ) ∂'i ( g' B 'i/h'i)} ^
en B = B'n ^
en
V = h'n [ (1/ g' ) ε'ncd ε'eab ∂'c {(1/ g' ) g'de(∂'a[g'bf B 'f/h'f]) }] ^
en . (13.1.19)

To this list we can add the Cartesian form

@B = [∇2Bn ] ^
n . // Cartesian, ^
n = un B = Bn ^
n (13.1.20)

Converting from Picture B to Picture MS using rules (9.3.2) gives,

(13.1.21)
@B = G – V

G = ∂n{ (1/ g ) ∂i ( g Bi)} en B = Bnen

V = [ (1/ g ) εncd εeab ∂c { (1/ g ) gde(∂a[gbfBf]) }] en

G = hn ∂n{ (1/ g ) ∂i ( g B i/hi)} ^

en B = Bn ^
en
V = hn [ (1/ g ) εncd εeab ∂c { (1/ g ) gde(∂a[gbfBf/hf]) }] ^
en

@B = [∇2Bn ] ^
n // Cartesian, ^
n = un B = Bn ^
n (13.1.22)

In the equations above, all the ∂i mean ∂/∂ui and the argument u of all functions is suppressed. The
Cartesian form will be verified below.

Comment: The above general statement of @B in curvilinear coordinates is amazingly complicated.

nc d
Using (D.10.37-39), one could replace εncd εeab = δ(ncd;eab) = δ e a b = δn,eδc,aδd,b + 5 similar terms,

194
 Chapter 13: Vector Laplacian

but this just expands the V term into six terms in place of one. The alternate path of starting with [@B]n =
B'n;j;j as shown in Section 15.8 leads to a result involving the affine connection Γ, as well as ∂iΓ and
ΓΓ terms and is also quite unpleasant. We must accept the fact that @B is a complicated object which
complicates the study of differential equations (such as Navier-Stokes) in general coordinates.

13.2 The Vector Laplacian in orthogonal curvilinear coordinates

We continue in Picture M&S and shall use only the ^ e n expansion. Setting gij = hi2δi,j and gij =
(1/hi)2δi,j from (5.11.8), our equation (13.1.22) simplifies somewhat,

@B = G – V = hn [gnk ∂k{ (1/ g ) ∂i ( g Bi(u)/hi)}

– (1/ g ) εncd εeab ∂c{ (1/ g ) gde (∂a [gbf Bf(u)/hf]) } ]^
en

= hn [δk,n ∂k{ (1/ g ) ∂i ( g Bi(u)/hi)} (1/hn)2

– (1/ g ) εncd εeab ∂c{ (1/ g ) hd2δd,e (∂a [hb2δb,f Bf(u)/hf]) } ] ^
en

= hn [ ∂n{ (1/ g ) ∂i ( g Bi(u)/hi)} (1/hn)2

– (1/ g ) εncd εdab ∂c{ (1/ g ) hd2 (∂a [hb Bb(u)]) } ] ^
en

= [ (1/hn) ∂n{ (1/ g ) ∂i ( g Bi/hi)} – (hn/ g ) εncd εdab ∂c{ (1/ g ) hd2 (∂a [hb Bb]) } ] ^
en .

= [ An – (hn/ g ) Bn ] ^
en (13.2.1)

where An ≡ (1/hn) ∂n{ (1/ g ) ∂i ( g Bi/hi)}

Bn ≡ εncd εdab ∂c{ (1/ g ) hd2 (∂a [hb Bb]) } . (13.2.2)

Now define,

T ≡ (1/ g ) ∂i ( g Bi/hi)
Γd ≡ (1/ g ) hd2 εdab{ (∂a [hb Bb]) } ) (13.2.3)

so that

An = (1/hn) ∂nT
Bn = εncd ∂cΓd . (13.2.4)

Then (13.2.1) says

@B = [ (1/hn) ∂nT – (hn/ g ) εncd ∂cΓd ] ^

en
where T ≡ (1/ g ) ∂i ( g Bi/hi) and
Γd ≡ (1/ g ) hd2 εdab{ (∂a [hb Bb]) } ) . (13.2.5)

195
 Chapter 13: Vector Laplacian

It is conventional to write this out more specifically as

@B = [ (1/h1) ∂1T – (h1/ g ) ε1cd ∂cΓd ] ^

e 1 + cyclic

But ε1cd ∂cΓd = (∂2Γ3 - ∂3Γ2) so

@B = [ (1/h1) ∂1T – (h1/ g ) (∂2Γ3 - ∂3Γ2)] ^

e 1 + cyclic

= [(1/h1) ∂1T + (h1/ g ) (∂3 Γ2 – ∂2 Γ3) ] ^e1

+ [(1/h2) ∂2T + (h2/ g ) (∂1 Γ3 – ∂3 Γ1) ] ^
e2
+ [(1/h3) ∂3T + (h3/ g ) (∂2 Γ1 – ∂1 Γ2) ] ^
e3 . // M&S 1.11 (13.2.6)

Writing out the Γd from (13.2.3),

Γ1 = (1/ g ) h12 ( ∂2 [h3 B3] – ∂3 [h2 B2] )

Γ2 = (1/ g ) h22 ( ∂3 [h1 B1] – ∂1 [h3 B3] )
Γ3 = (1/ g ) h32 ( ∂1 [h2 B2] – ∂2 [h1 B1] ) .

T = (1/ g ) ∂i ( g Bi/hi) (13.2.7)

With the replacements

B→E Bn→ En ^
e n→ a n hi → gii T→ϒ (13.2.8)

equations (13.2.6) and (13.2.7) agree with Moon and Spencer p 3 (1.11).

13.3 The Vector Laplacian in Cartesian coordinates

First we verify that our general orthogonal coordinates form (13.2.5) reduces to the expected result for the
case that u-space is the same Cartesian space as x-space. One then has,

hi = 1 g = 1 u = F(x) = x ⇒ R=S=1
(^ i i
e n) = Sn = δn i
⇒ ^ en = ^
n,
B = Bn ^
e n = Bn ^
n ⇒ Bn = Bn = Bn . (13.3.1)

Then (13.2.5) becomes

@B = [ ∂nT – εnab∂a Γb ] ^
n // implied sum on n
i
where T = ∂iB
Γb = εbcd (∂cBd) (13.3.2)
or

196
 Chapter 13: Vector Laplacian

@B = [ ∂n{∂iBi} – εnab∂a { εbcd (∂cBd)} ] ^

n
nab ^
= [ ∂n{div B} – ε ∂a { (curl B)b } ] n
= [ ∂n{div B} – [curl (curl B)]n } ] ^
n
= grad(div B) – curl (curl B) (13.3.3)

which agrees with our starting definition of @B (13.1.4).

Second, we verify the claim of (13.1.3) that in Cartesian coordinates

[@B]n = ∇2 Bn . (13.1.3)

We start with the first line of (13.3.3),

[@B]n = ∂n∂iBi – εnab εbcd ∂a∂cBd

= ∂n∂iBi – (δncδad – δndδac) ∂a∂cBd // from (D.10.22)

= ∂n∂iBi – ∂n∂dBd + ∂c∂cBn

= ∂2cBn = ∇2(Bn) = ∇2(Bn) . (13.3.4)

Comment: This is a proof of the Cartesian-coordinates vector identity [∇(∇•B) - ∇x(∇xB)]n = ∇2(Bn) .

Thus it has been shown that, in Cartesian coordinates,

[@B]n = ∇2 (Bn) = [ grad(div B) – curl (curl B) ]n . (13.3.5)

It is this second and rather complicated form which allowed us to obtain the expression (13.2.5) for @B
expressed in general curvilinear coordinates.

197
 Chapter 14: Summary

14. Summary of Differential Operators in curvilinear coordinates

14.1 Summary of Conventions and How To

The results are given in the Picture M&S context, and are copied from Chapters 9-13.

(14.1.1)

The Standard Notation of Chapter 7 is used throughout the tables.

In all the differential operator equations below, an operator acts either on a tensorial vector field B or on a
tensorial scalar field f. On the right side of the drawing above, objects are said to be in x-space, and f(x)
and Bn(x) (components of B) are x-space tensorial objects. On the left side of the drawing objects are said
to be in u-space. The function f is represented in u-space as f (u), while there are three different ways to
represent the components of B, called Bn(u), Bn(u) and Bn(u). There is a big distinction between the x-
space objects and the u-space objects. For the scalar, f (u) = f(x) = f(x(u)) and so f has a different
functional form than f. For the vector components, the u-space components are linear combinations of the
x-space components, for example Bn = RnmBm ( or B = RB, contravariant vector transformation).

See (9.3.6) concerning the font used for Bn(u).

On lines marked "Cartesian", ∂n = ∂/∂xn and f(x) and Bn(x) appear (Cartesian components).

On other lines, ∂n = ∂/∂un, and the f and B objects appear in italics and are functions of u. The other
functions like hn, gab and g are also functions of u.

The vectors en, ^e n and en all exist in Cartesian x-space. The en are the tangent base vectors of Chapter 3,
and the en are the reciprocal base vectors of Chapter 6. The unit vectors ^ e n ≡ en/ |en| = en/hn are used as
well.

The dot product A•B is the covariant one of Section 5.10, namely, AiB̄i → AiBi = gijAiBj .

For each differential operator, the object on the LHS of the equations is always the same: it is a
differential operator acting on f(x) or B(x) in x-space. In the Cartesian lines, the RHS expresses that LHS
object in terms of Cartesian objects and Cartesian coordinates. On the other lines, the RHS expresses that
exact same LHS x-space object in terms of Curvilinear (u-space) objects and coordinates. When the LHS
is a scalar, the LHS object can be considered to be in either x-space or u-space. When the LHS is a vector,
that LHS object is in x-space but can be related to u-space objects by a linear transformation by R.

198
 Chapter 14: Summary

How to Compute Things for some Arbitrary Coordinate System

For some arbitrary (and possibly non-orthogonal) coordinate system in some arbitrary number of
dimensions N, one can start with the defining equations such as the following ones for spherical
coordinates,

x = rsinθcosφ that is: x = F-1(u ) u = (r,θ,φ)

y = rsinθsinφ
z = rcosθ .

In the Picture M&S context and in developmental notation, one uses (2.1.5) Sik ≡ (∂xi/∂uk) to compute
the matrix S. Then the covariant metric tensor comes from (5.3.3), ḡ = STGS. Normally one has G = 1
for Cartesian x-space, but see (1.10) for quasi-Cartesian. The scale factors are then hi = ḡ ii and g =
det(ḡ), as shown in (5.11.7) and (5.12.20). The contravariant metric tensor is then found as g = ḡ -1
(matrices). The conversion to Standard Notation is then ḡ→ gab and g→gab. These calculations are easily
automated as shown for example in the Maple code of (G.6.1) and (G.6.2).
If the coordinate system is orthogonal, then ḡ will be diagonal and one will have ḡ nm = hn2 δn,m and
gnm = hn-2 δn,m and finally g = det(ḡ) = Πihi. For the eleven classical orthogonal systems the scale
factors hi are easily found online or in other reference sources, so one can then avoid the above steps (or
verify them).
At this point, any formulae appearing in the tables below can be used.

199
 Chapter 14: Summary

The expressions marked below appear on pages 2 or 3 of Moon & Spencer (M&S).

general: g ≡ det(gab) hn2 ≡ gnn ∂i = gij∂j Bn = hnBn en = hn^

en
orthogonal: g = (Πihi) = h1h2...hN gnm = hn2 δn,m gnm = hn-2 δn,m (14.1.2)
___________________________________________________________________________

14.2 divergence

divergence general: (9.3.3) (14.2.1)

[div B](x) = [1/ g ] ∂n [ g Bn] B = Bnen

[div B](x) = [1/ g ] ∂n [ g Bn/ hn ] B = Bn^
en // M&S 1.06
nm n
[div B](x) = [1/ g ] ∂n [ g g Bm] B = Bn e
[div B](x) = ∂nB n n
// B = Cartesian components of B B = Bn ^n= Bn ^
n
[div B](x) = [div B](u) // transformation (scalar)

divergence orthogonal: (9.3.4) (14.2.2)

[div B](x) = [1/(Πihi)] ∂n [(Πihi) Bn] B = Bnen

[div B](x) = [1/(Πihi)] ∂n [(Πihi) Bn / hn] B = Bn ^
en
[div B](x) = [1/(Πihi)] ∂n [(Πihi) Bn / hn2] B = Bnen

divergence orthogonal N=3: (9.3.5) (14.2.3)

[div B](x) = [1/(h1h2h3)] { ∂1[h2h3 B1(u) ] + cyclic } B = Bn ^

en
___________________________________________________________________________

14.3 gradient and gradient dot vector

gradient general: (10.1.16) (14.3.1)

[grad f](x) = (∂if ) ei

[grad f](x) = (∂if ) ei = gij (∂jf ) ei
[grad f](x) = hi(∂if ) ^
ei = hi gij (∂jf ) ^
ei
^
[grad f](x) = (∂if) n // Cartesian
j
[grad f ]i(u) = Ri [grad f]j(x) // transformation (vector) f (u) = f(x) = f(x(u))

gradient orthogonal: (10.1.17) (14.3.2)

[grad f](x) = (∂if ) ei

[grad f](x) = (∂if ) ei = (1/hi2) (∂if ) ei
[grad f](x) = hi(∂if ) ^
ei = (1/hi) (∂if ) ^
ei // M&S 1.05

200
 Chapter 14: Summary

gradient dotted with a vector: (10.2.5) (14.3.3)

[grad f](x) • B(x) = (∂nf ) Bn B = Bnen

[grad f](x) • B(x) = (∂nf ) Bn B = Bnen
[grad f](x) • B(x) = (∂nf ) Bn/hn B = Bn ^
en
[grad f](x) • B(x) = (∂nf) Bn // Cartesian B = Bn ^
n
[grad f](x) • B(x) = [grad f ](u) • B(u) // transformation (scalar)
___________________________________________________________________________

14.4 Laplacian

Laplacian general: (11.10) (14.4.1)

[lap f](x) = [1/ g ] ∂m[ g gnm (∂nf ) ]

[lap f](x) = ∂n2f(x) // Cartesian f (u) = f(x) = f(x(u))
[lap f](x) = [lap f ](u) // transformation (scalar)

Laplacian orthogonal: (11.12) (14.4.2)

[lap f](x) = [1/(Πihi)] ∂m[ (Πihi) (1/hm2) (∂mf ) ] // orthogonal // M&S 1.09

Laplacian orthogonal N=3: (11.13) (14.4.3)

[lap f](x) = 1/(h1h2h3) { ∂1 [ (h2h3/h1) ∂1f ] + cyclic }

___________________________________________________________________________

201
 Chapter 14: Summary

14.5 curl

curl general: (12.4.7-11) // N=3 only (14.5.1)

[curl B](x) = εnab [(1/ g ) ∂aBb ] en B = Bnen

[curl B](x) = εnab [(1/ g ) ∂a(gbcBc )] en B = Bnen
[curl B](x) = εnab[(1/ g ) hn ∂a(gbc Bc/hc )] ^
en B = Bn^
en
[curl B](x) = ε nab
∂aBb(x) ^
n // Cartesian B = Bn ^
n

⎪ e1 e2 e3 ⎪
[curl B](x) = (1/ g ) ⎪ ∂1 ∂2 ∂3 ⎪ B = Bnen
⎪ ⎪
⎪ B1 B2 B3 ⎪

⎪ e1 e2 e3 ⎪
[curl B](x) = (1/ g ) ⎪ ∂1 ∂2 ∂3
⎪ B = Bnen
⎪ c c c ⎪
⎪ g1cB g2cB g3cB ⎪

⎪ h1 ^e1 h2 ^e2 h3 ^e3 ⎪

[curl B](x) = (1/ g ) ⎪ ∂1 ∂2 ∂3 ⎪ B = Bn^
en // M&S 1.07
⎪ c c c ⎪
⎪ (g1c/hc) B (g2c/hc) B (g3c/hc) B ⎪

⎪ ^
1 ^ 2 ^3 ⎪
[curl B](x) = ⎪ ∂1 ∂2 ∂3 ⎪ // here ∂n= ∂/∂xn and Bi = Bi(x) B = Bn ^
n
⎪ ⎪
⎪ B1 B2 B3 ⎪

curl orthogonal: (12.5.2-5) (14.5.2)

[curl B](x) = εnab [(h1h2h3)-1 ∂aBb ] en B = Bnen

[curl B](x) = εnab [(h1h2h3)-1 ∂a(hb2Bb )] en B = Bnen
[curl B](x) = εnab[(h1h2h3)-1 hn ∂a(hb Bb)] ^ en B = Bn^
en

⎪ e1 e2 e3 ⎪
[curl B](x) = (h1h2h3)-1 ⎪ ∂1 ∂2 ∂3 ⎪ B = Bnen
⎪ ⎪
⎪ B1 B2 B3 ⎪

⎪ e1 e2 e3 ⎪
[curl B](x) = (h1h2h3)-1 ⎪ ∂1 ∂2 ∂3 ⎪ B = Bnen
⎪ 2 1 2 2 2 3 ⎪
⎪ 1
h B h 2 B h 3 B ⎪

⎪ h1 ^e 1 h2 ^e 2 h3 ^e 3 ⎪
-1⎪
[curl B](x) = (h1h2h3) ∂1 ∂2 ∂3 ⎪ B = Bn^
en // M&S 1.07a
⎪ 1 2 3 ⎪
⎪ h1 B h2 B h3 B ⎪
___________________________________________________________________________

202
 Chapter 14: Summary

14.6 vector Laplacian

vector Laplacian general: (13.1.22) // N=3 only (14.6.1)

[@ B](x) = G – V

G = ∂n{ (1/ g ) ∂i ( g Bi)} en B = Bnen

V = [ (1/ g ) εncd εeab ∂c { (1/ g ) gde(∂a[gbfBf]) }] en

G = hn ∂n{ (1/ g ) ∂i ( g B i/hi)} ^

en B = Bn ^
en
V = hn [ (1/ g ) εncd εeab ∂c { (1/ g ) gde(∂a[gbfBf/hf]) }] ^
en

@ B = [∇2Bn ] ^
n // Cartesian, ^
n = un B = Bn ^
n

vector Laplacian orthogonal: (13.2.6-7) (14.6.2)

[@ B](x) = [(1/h1) ∂1T + (h1/ g ) (∂3 Γ2 – ∂2 Γ3) ] ^

e1 // M&S 1.11
+ [(1/h2) ∂2T + (h2/ g ) (∂1 Γ3 – ∂3 Γ1) ] ^
e2
+ [(1/h3) ∂3T + (h3/ g ) (∂2 Γ1 – ∂1 Γ2) ] ^
e3

T = (1/ g ) ∂i ( g Bi/hi)
Γ1 = (1/ g ) h12 ( ∂2 [h3 B3] – ∂3 [h2 B2] )
Γ2 = (1/ g ) h22 ( ∂3 [h1 B1] – ∂1 [h3 B3] )
Γ3 = (1/ g ) h32 ( ∂1 [h2 B2] – ∂2 [h1 B1] )

or (13.2.5) (14.6.3)

[@ B](x) = [ (1/hn) ∂nT – (hn/ g ) εnab∂a Γb ] ^

en

T = (1/ g ) ∂i ( g Bi/hi)
Γb = (1/ g ) hb2 εbcd ( ∂c [hd Bd])
___________________________________________________________________________

203
 Chapter 14: Summary

14.7 Example 1: Polar coordinates: a practical curvilinear notation

From our many visits to this example we know that :

general: g ≡ det(gab) hn2 ≡ gnn ∂i = gij∂j Bn = hnBn en = hn^

en
orthogonal: g = (Πihi) = h1h2...hN gnm = hn2 δn,m gnm = hn-2 δn,m (14.1.2)

e1 = r(-sinθ,cosθ) = eθ = r ^
eθ // = r ^
θ
e2 = (cosθ,sinθ) = er = ^
er // = ^ r (3.4.2) (14.7.1)

θ r x y
2
r 0⎞θ -sinθ/r cosθ/r θ
gij = ⎛ R = ⎛⎝ cos(θ) sinθ ⎞⎠ r u1 = θ u2 = r (!!) (14.7.2)
⎝ 0 1⎠r
(5.13.11) (3.4.1)

h 1 = hθ = gθθ = r h 2 = hr = grr = 1 . (5.13.11) (14.7.3)

Assume one is working with a 2D vector velocity field v(x) -- our first encounter with a "lower case"
vector field which we have been careful to support with the general notations above. Since one knows the
names of the variables 1=θ and 2=r, one might define the following new variables on the first line to be
the officially named variables on the second line

vθ vr vθ vr vθ vr vx vy
v1 v2 v1 v2 v 1
v 2
v1=v 1
v2=v2
contravariant covariant unit vector Cartesian
(italic) (italic) (non-italic) (non-italic) (14.7.4)

One ends up with the comfortable v = vθ^

θ + vr ^
r notation as shown below.

vθ ≡ v1 = h1ν1 = hθ vθ = r vθ // unit vector projection components

vr ≡ v2 = h2ν2 = hr vr = vr

vθ = Rθxvx + Rθyvy = -sinθ/r vx + cosθ/r vy

vr = Rrxvx + Rryvy = cosθ vx +sinθ vy (14.7.5)
so
vθ = r vθ = -sinθ vx + cosθ vy
vr = vr = cosθ vx +sinθ vy

v = vnen = vθeθ + vrer

v = vn^ e n = vθ^
e θ + vr^e r = vθ^
θ + vr ^
r
^ ^ ^
v = v n n = vx x + v y y . (14.7.6)

As an example of a differential operator, consider the divergence for orthogonal coordinates from (14.2.2)

204
 Chapter 14: Summary

[div B](x) = [1/(Πihi)] ∂n [(Πihi) Bn / hn] . B = Bn ^

en (14.7.7)

Applied to the present situation one gets (h1 = hθ = r and h2 = hr = 1),

[div v](x) = [1/(h1h2)] { ∂1 [h2 v1] + ∂2[h1 v2] } V = vn ^

en
= [1/(hθhr)] { ∂θ [hr vθ ] + ∂r[hθ vr] }
= (1/r) { ∂θvθ + ∂r(rvr) } . (14.7.8)

Suppose vx and vy are constants. Then the Cartesian expression says

[div v](x) = ∂nvn = ∂xvx + ∂yvy = 0 + 0 = 0 . (14.7.9)

Equation (14.7.9) agrees :

[div v](x) = (1/r) { ∂θvθ + ∂r(rvr) }

= (1/r) { ∂θ[-sinθ vx + cosθ vy] + ∂r(r [cosθ vx + sinθ vy]) }
= (1/r) { [-cosθ vx - sinθ vy] + [cosθ vx + sinθ vy] }
=0. (14.7.10)

The notation illustrated here works for any curvilinear coordinates.

205
 Chapter 15: Covariant Method

15. Covariant derivation of all curvilinear differential operator expressions

15.1 Review of Chapters 9 through 13

Let us briefly review the discussion of Chapters 9 through 13 concerning the expression of differential
operators

div, grad, lap, curl and @

in curvilinear coordinates. The underlying framework was provided by Picture B,

(15.1.1)

where the coordinates x'n of x'-space were the curvilinear coordinates of interest, while the coordinates of
x-space were the Cartesian coordinates. The transformation x' = F(x) provided the connection between the
curvilinear coordinates and the Cartesian ones.

In Chapter 9 the object div B was treated as the total B "flux" (∫B•dA) emitting from an N-piped in x-
space divided by the volume of that N-piped, in the limit the N-piped shrank to a point. Using Appendix
B formulas for the N-piped area, and using the key expansion

B = B'nen , // (7.13.10) line 3

where the en are tangent base vectors in x-space and the B'n are components of B in x'-space, we obtained
in (9.1.18) a way to write the Cartesian-space div B in terms of curvilinear x' coordinates and x'-space
objects, namely, B'n(x'), g'(x') and the gradient operator ∂'i = ∂/∂x'i.

In Chapter 10 the vector grad f was expanded using (7.13.10) line 4 that V = V'nen to get

grad f = [grad f]'nen = (∂'nf ')en

where en are the reciprocal base vectors. This result was expressed in various ways, and gradf •B was
also treated such that gradf •B = (∂'nf ')en • B'mem = (∂'nf ') B'n . [ Recall that f '(x') = f(x). ]

In Chapter 11 the Laplacian was written as lap f = div (grad f) and then the results of Chapters 9 and 10
for div and grad were used to obtain (11.5) which expresses lap f in terms of x'-space coordinates and
objects B'n(x'), g'(x'), g'nm(x') and ∂'i . In traditional notation, one writes ∇2f = ∇ • [ ∇f ] = div (grad f).

In Chapter 12 the x'-space curl component C'n = [curl B]'n was treated as the circulation line integral of
B around near face n of the same N-piped used in Chapter 9 for divergence, divided by the area of that

206
 Chapter 15: Covariant Method

face, again in the limit the N-piped shrank to a point. With the expansions B = B'nen and C = J-1C'nen,
this led to an expression (12.3.3) for curl B in terms of the curvilinear x' coordinates and the x'-space
objects B'n(x'), g'(x'), g'nm(x'), ε'nab and ∂'i .

In Chapter 13 the vector Laplacian was treated using the definition

@ B ≡ grad(div B) – curl (curl B)

and then the results of Chapters 9, 10 and 12 on div, grad and curl were recruited to produce an
expression (13.1.17) for @ B in terms of x'-space coordinates and various x'-space objects.

15.2 The Covariant Method

We shall now repeat all of the above work using "the covariant method" which, as will be shown, gets
most results extremely quickly, but at a cost of requiring knowledge of tensor densities and covariant
differentiation, as described in Appendices D and F.

Imagine that one has a Cartesian x-space expression for some tensor object of interest Q,

Q---- = [ Cartesian form with various up and down indices and various derivatives ] (15.2.1)

where Q---- means that Q has some number of up and down indices. The following four steps should then
be carried out:

1. Arrange any summed index pairs within [....] so they are "tilted". Then when [...] is "tensorized" those
tilted pairs will be true contractions.
Note that doing this does not affect the Cartesian value of this object [...] since up and down indices
are the same in Cartesian space.

2. Replace all derivatives with covariant derivatives. Thus is done by first writing all derivatives in
comma notation, and then by replacing those commas with semicolons. For example:

∂αBa = Ba,α → Ba;α . // Ba;α = ∂α Ba – ΓnaαBn (F.9.2) (15.2.2)

Objects like ∂αBa which are not tensors become objects like Ba;α which are tensors according to the
theorem of (F.7.1).
Note again that doing this does not affect the Cartesian value of this object [...], because in Cartesian
x-space Γnaα = 0 as per (F.4.16). For a more general case like Babc..x;α shown in (F.7.2), making the
replacement Babc..x,α makes no difference as long as Babc..x is a true tensor (W = 0), because all those
Γ correction terms seen in (F.7.2) vanish in Cartesian space. The same is true regardless of index
positions.

3. Insert appropriate powers of g as needed to achieve appropriate tensor density weights so that all terms
in [...] have the same weight and this weight equals the weight of Q---- . Tensor densities are discussed in
Appendix D, and this weight adjustment idea appears in (D.2.3c).

207
 Chapter 15: Covariant Method

Note again that doing this does not affect the Cartesian value of this object [...] because g = 1 in
Cartesian space (g = det(gij) ).

Comment: Since x-space is Cartesian, we know from (5.12.16) that g > 0 and g' > 0, so we don't need to
use expressions like |g| and |g'| in our current context.

4. At this point one has

Q---- = { tensorized form of the Cartesian expression of interest } . (15.2.3)

Since this is then a "true tensor equation" (it could be a true tensor density equation), according to the rule
of covariance described in Section 7.15 the above equation can be converted to x'-space by simply
priming Q and all objects within { }. One then has

Q' ---- = { tensorized form of the Cartesian expression of interest }' (15.2.4)

where the notation {...}' means that everything inside {...} is primed.

Perhaps at this point one does some simplifications of the resulting {...}' object. The result then is the
expression of Q---- in general curvilinear coordinates if the underlying transformation x' = F(x) describes
the transformation from Cartesian to curvilinear coordinates. For a vector, recall (7.13.10) that Q = Q'nen
so the "curvilinear coordinates" expression of Q involves the the x'-space components Q'a(x') but the
tangent base vectors en are in x-space. A similar statement can be made for Q being any tensor, as shown
for example in (E.2.9) which says Q = Σijk Q'ijk (ei⊗ej⊗ek) .

These four steps then comprise "the covariant method". The method of course applies to tensor analysis
applications other than "curvilinear coordinates".

Uniqueness. One might obtain two tensorizations of a Cartesian expression that look different. Both these
tensorizations must lead to the same Q'---- in x'-space. For example, suppose one finds two tensorizations
of a vector Qa , call them {...}1a and {...}2a which, when evaluated in Cartesian x-space, are equal

{...}1a = {...}2a when both sides are evaluated in Cartesian x-space . (15.2.5)

One can then transform both objects to x'-space in the usual manner,

{...}'1a = Rab{...}1b and {...}'2a = Rab{...}2b . (15.2.6)

Therefore, since {...}1a = {...}2a in x-space, one must have {...}'1a = {...}'2a in x'-space.

Example: In (15.7.6) below we will establish the following tensorization of (@B)n (the vector Laplacian),

(@B)n = {.....}1n = (Bj;j);n – g-1/2εnab(g-1/2gbcεcdeBe;d);a . (15.2.7)

208
 Chapter 15: Covariant Method

Since in Cartesian x-space one can write (@B)n = ∇2(Bn) = ∂j∂jBn = Bn,j,j , another tensorization of
(@B)n is given by

(@B)n = {.....}2n = Bn;j;j . (15.2.8)

Therefore, the following must be true in x'-space

B'n;j;j = (B'j;j);n – g'-1/2ε'nab(g'-1/2g'bcε'cdeB'e;d);a (15.2.9)

Explicitly verifying (15.2.9) takes some work and we do it below in Section 15.8. Both forms are used in
Appendix I to compute @B in spherical and cylindrical coordinates.

In the following Sections we shall in short order derive the curvilinear expressions for the basic
differential operators using the covariant method outlined above. This method is used as well (along with
parallel brute force methods) to obtain expressions for objects ∇v, divT,∇T in Appendices G, H, J.

15.3 divergence (Chapter 9)

In Cartesian coordinates, [div B] = ∂iBi = Bi,i. The true tensorial tensor which matches this is [div B] ≡
Bi;i since Γ = 0 for Cartesian coordinates by (F.4.16). That is to say, from (F.9.3),

Ba;α = ∂α Ba + Γaαn Bn (F.9.3)

so
Bi;i = ∂i Bi + Γiin Bn = ∂i Bi = Bi,i since Γcab ≡ 0 .

Since Bi;i is a scalar, it is the same in x-space as in x'-space (Section 7.15 covariance),

[div B] = [div B]' = B'i;i = ∂'i B'i + Γ 'iin B'n (15.3.1)

where B'n are the contravariant components of vector B in x'-space. But recall (F.4.2) in x'-space,

Γ'iin = (1/ g' ) ∂'n( g' ) . (F.4.2)

Therefore

[div B] = ∂'n B'n + (1/ g' ) ∂'n( g' )B'n = (1/ g' ) ∂'n ( g' B'n). (15.3.2)

This matches the result (9.2.1) obtained in Chapter 9 by geometric methods.

15.4 gradient and gradient dot vector (Chapter 10)

If f is a scalar, then G = grad f transforms as an ordinary vector according to (2.4.2) and (2.5.1), so it is
already a tensor. According to line 4 of (7.13.10) the expansion of such a vector may be written

209
 Chapter 15: Covariant Method

G = [grad f] = G'iei = (∂'if ') ei . (15.4.1)

Also, grad f • B is a scalar, again already a tensor. Using this fact and line 3 of (7.13.10) that B = B'nen ,

[grad f]' • B' = [grad f] • B = [(∂'if ') ei] • (B'nen) = (∂'if ')B'nδin = (∂'if')B'i (15.4.2)

These two results match (10.1.7) and the second line of (10.2.4) of Chapter 10. As a reminder, since f is a
scalar field, f '(x') = f(x).

15.5 Laplacian (Chapter 11)

In Cartesian x-space the Laplacian of a scalar function f can be written as

[lap f]Cart = ∂i∂if = f,i,i . (15.5.1)

The tensorized version of the Laplacian is then the following, using the ,→; rule,

lap f = f;i;i . (15.5.2)

To verify that this is so (we already know it is so), we call upon (F.9.19) with scalar B → f,

f;a;α = ∂α f;a + Γaαn f;n . (F.9.19) (15.5.3)

Then (F.4.16) says Γ = 0 so f;a;α = ∂α f;a = ∂α f,a = f,a,α so finally f;i;i = f,i,i .

Since (15.5.3) is a true tensor equation, Section 7.15 covariance says it takes this form in x'-space,

f';a;α = ∂'α f';a + Γ'aαn f';n . (15.5.4)

But lap f = f;i;i is a scalar, so we have

[lap f] = [lap f]' = f';i;i = ∂'i f';i + Γ'iin f';n = ∂'i f',i + Γ'iin f',n

= ∂'i∂'if' + Γ'iin (∂'n f') . (15.5.5)

But identity (F.4.2) says Γ'iin = (1/ g' ) ∂'n( g' ) so then

lap f = ∂'i∂'if' + (1/ g' ) ∂'n( g' ) (∂'n f')

= [1/ g' ] ∂'n [ g' (∂'nf ')] . (15.5.6)

This matches (11.5) found in Chapter 11.

210
 Chapter 15: Covariant Method

15.6 curl (Chapter 12)

In Cartesian space we know that C = curl B and Ci = εijk∂jBk = εijk Bk,j. The obvious tensorized
form is Ci = εijk Bk;j which we could verify using (F.9.2) with Γ = 0, but we know this is correct from
the general theory (Step 2) presented in Section 15.2. Since this is a true tensor equation, Section 7.15
covariance tells us that it takes this form in x'-space,

C'i = ε'ijk B'k;j . (15.6.1)

Recall from (D.4.8) that ε'ijk = εijk = the permutation tensor.

As shown in Section D.8, C ≡ curl B is a vector density of weight -1 because it contains the ε tensor
which has that weight. According to (E.2.26) the expansion of such a vector density is given by
C = J-1C'iei . Therefore,

curl B = C = J-1C'i ei = J-1 [ε'ijkB'k;j]ei = (1/ g' ) ε'ijk B'k;jei (15.6.2)

But we now do call upon (F.9.2) in x'-space to show that ε'ijk B'k;j = ε'ijk B'k,j :

B'a;α = ∂'α B'a – Γ'naαB'n covariant rank-2 tensor // 2nd term is sym on a↔α (F.9.2)
so
B'k;j = ∂'j B'k – Γ'nkjB'n
so
ε'ijk B'k;j = ε'ijk [∂'j B'k – Γ'nkjB'n] = ε'ijk ∂'j B'k – ε'ijkΓ'nkjB'n

= ε'ijk ∂'j B'k – 0 // because ε' is antisymmetric on jk, but Γ'nkj is symmetric on jk †

= ε'ijk B'k,j . (15.6.3)

Inserting this result into (15.6.2) gives

curl B = (1/ g' ) ε'ijk B'k,jei = (1/ g' ) ε'ijk(∂'jB'k)ei .

This matches the result (12.3.3) found in Chapter 12 based on geometric circulation integrals.

†
Q ≡ AijSij = AjiSji = (-Aij)(+Sij) = - AijSij = -Q ⇒ Q=0 (15.6.4)
i↔j symmetries

15.7 vector Laplacian (Chapter 13)

The vector Laplacian operator takes a bit more work. Since this is a major example of carrying out the
"tensorization process" of Section 15.2, much detail is provided.

In Chapter 13 the vector Laplacian is written @ (instead of ∇2) and is defined by

211
 Chapter 15: Covariant Method

@B ≡ grad(div B) – curl (curl B) . (13.1.4) (15.7.1)

Recall from (13.3.4) that that in Cartesian space,

[grad(div B) – curl (curl B)]n = ∇2(Bn) = (@B)n // Cartesian x-space (15.7.2)

so we are allowed to use the LHS here to find the curvilinear expression of @B . In components,

(@B)n = ∂n(∂jBj) - εnab∂a [curl B]b

= ∂n(∂jBj) - εnab∂a (εbde∂dBe) . (15.7.3)

In Cartesian space up and down index position does not matter, see (5.9.1), and we are just jockeying the
indices in search of a tensorized form with contracted indices where possible. Writing (15.7.3) in comma
notation,

(@B)n = (Bj,j),n – εnab(εbdeBe,d),a (15.7.4)

one is then free to replace commas with semicolons since Γ = 0 in Cartesian coordinates, see (F.4.16),

(@B)n = (Bj;j);n – εnab(εbdeBe;d);a . (15.7.5)

There is a "technical difficulty" visible here. The first term (Bj;j);n is a true vector since it is the
contraction of a true rank-3 tensor on two tilted indices, but the second term is a rank-3 tensor density of
weight -2. In the second term, in order to neutralize the weight of each ε tensor density (from (D.4.9) each
ε has weight -1), we shall now add a benign factor of g-1/2, as suggested in (D.2.3c), since g-1/2 is a
scalar density of weight +1. This in turn is so from (5.12.13) that g' = J2g which says g has a weight of -2.
The factor g-1/2 is benign since g = 1 in Cartesian coordinates. So, with this upgrade installed,

(@B)n = (Bj;j);n – g-1/2εnab(g-1/2εbdeBe;d);a . (15.7.6)

Now both objects Tb ≡ (g-1/2εbdeBe;d) and Vn ≡ g-1/2εnab Tb;a are true vectors, so the above equation
says that a true vector is the difference between two other true vectors.

Anticipating a few steps ahead when our ship lands in x'-space, we know that εbdeBe;d = εbdeBe,d due
to the e↔d symmetry of the Γ term in Be;d = Be,d - ΓsedBs of (F.9.24), and this motivates us to perform
a "tilt change" in this product. Such a tilt-change also gets the ;d index "down" which is a nicer place for
it to be since it will soon become ∂d. The tilt operation is done as follows ,

εbdeBe;d = εbdeBe;d = εbdeBe;d = gbc εcdeBe;d . (15.7.7)

Recall that tilt changes like this are only allowed with true tensor indices, (7.11.3), and that is why the ε
tilt reversal can be done only after the semicolon in Be;d is installed. Eq. (15.7.6) then reads,

212
 Chapter 15: Covariant Method

(@B)n = (Bj;j);n – g-1/2εnab(g-1/2gbcεcdeBe;d);a . (15.7.8)

Although εbde is a fine tensor, we really prefer εcde since this is the permutation tensor, and that is why
the gbc factor is added in (15.7.7).

At this point every index is a tensor index, so (15.7.8) is a "true tensor equation" in the sense of Section
7.15. The above equation agrees exactly with the Cartesian expression of (@B)n since g = 1, gbc = δb,c,
and the semicolons are commas since Γ = 0, a fact verified in the Comment above. The point is that we
have found a way to write the Cartesian expression of (@B)n such that both terms are true tensors. The
equation is therefore now "covariant" and the equation in x'-space can be obtained simply by priming all
objects:

(@B)'n = (B'j;j);n – g'-1/2ε'nab(g'-1/2g'bcε'cdeB'e;d);a . (15.7.9)

The semicolons were installed to obtain a true tensor equation, but now that we have successfully arrived
in x'-space, we want to remove as many semicolons as possible because they imply "extra terms" since
now Γ' ≠ 0. Consider for example, this grouping which appears in the above,

ε'nab (g'-1/2g'bcε'cdeB'e;d);a = ε'nab T'b;a T'b ≡ (g'-1/2g'bcε'cdeB'e;d) . (15.7.10)

Since Tb;a = Tb,a – ΓnbaTn from (F.9.24), ε'nabT'b;a = ε'nab T'b,a because εnab is antisymmetric on a,b
while Γnba is symmetric, see (A.4.4). Thus,

(@B)'n = (B'j;j);n – g'-1/2ε'nab(g'-1/2g'bcε'cdeB'e;d),a . (15.7.11)

↑
cde cde
But ε' B'e;d = ε' B'e,d for the exact same symmetry reason, so

(@B)'n = (B'j;j);n – g'-1/2ε'nab(g'-1/2g'bcε'cdeB'e,d),a . (15.7.12)

↑
Meanwhile, (F.9.1) says D;n = D,n so that (B'j;j);n = (B'j;j),n , since the object B'j;j = B;j;j is a
scalar like D. Therefore,

(@B)'n = (B'j;j),n – g'-1/2ε'nab(g'-1/2g'bcε'cdeB'e,d),a . (15.7.13)

↑
cde
Since ε' is a constant (the permutation tensor, see below (D.4.1)), we now pull it through the ∂'a
derivative implied by ,a and then showing all derivatives the above becomes

(@B)'n = ∂'n(B'j;j) – g'-1/2ε'nabε'cde ∂'a(g'-1/2g'bc∂'dB'e) . (15.7.14)

From (15.3.2) and (F.9.1) that B;j = B,j we know that

[div B] = B;j;j = B,j;j = (1/ g' ) ∂'j ( g' B'j) = (1/ g' ) ∂'i ( g' B'i) . (15.7.15)

Using this in the first term on right of (15.7.14) gives,

213
 Chapter 15: Covariant Method

(@B)'n = ∂'n{(1/ g' ) ∂'i( g' B'i)} – g'-1/2ε'nabε'cde ∂'a(g'-1/2g'bc∂'dB'e) . (15.7.16)

To compare this result to that of Chapter 13, we shuffle the indices in the second term as follows,

g'-1/2ε'nabε'cde ∂'a(g'-1/2g'bc∂'dB'e) // now take a,b,c,d,e → A,B,C,D,E

g'-1/2ε'nABε'CDE ∂'A(g'-1/2g'BC∂'DB'E) // now take A→c, B→d, C→e, D→a, E→b
g'-1/2ε'ncdε'eab ∂'c(g'-1/2g'de∂'aB'b) . (15.7.17)

Then (15.7.16) reads,

(@B)'n = ∂'n{(1/ g' ) ∂'i( g' B'i)} – g'-1/2ε'ncdε'eab ∂'c(g'-1/2g'de∂'aB'b) . (15.7.18)

This may now be compared with (13.1.17) which we quote :

@B =G–V
G = ∂'n{ (1/ g' ) ∂'i ( g' B'i)} en B = B'nen
V = [ (1/ g' ) ε'ncd ε'eab ∂'c { (1/ g' ) g'de(∂'aB'b) }] en B = B'nen (15.7.19)

The results agree, and further processing of this result is done in Chapter 13.

This concludes our discussion of how the curvilinear-coordinate differential operator forms of Chapters
9,10,11,12,and 13 may be obtained relatively quickly using the notions of tensor densities (App. D) and
covariant derivatives (App. F), without resorting to the use of N-piped geometric constructs.

15.8 Verification that two tensorizations are the same

In Section 15.7 we obtained the x-space @B "tensorization" (15.7.8) which written in x'-space says

(@B)'n = B'j;j;n – g'-1/2ε'nab(g'-1/2g'bcε'cdeB'e;d);a . (15.7.9) (15.8.1)

Since in Cartesian coordinates (@B)n = ∇2(Bn) = ∂j∂j Bn = Bn,j,j , an obvious alternate tensorization
is Bn,j,j → Bn;j;j so that in x'-space

(@B)'n = B'n;j;j . (15.8.2)

In Section 15.2 we showed that tensorization is unique, so these two tensorizations must be the same. We
know they are the same, but in the time-honored tradition of "trust but verify", we would like to actually
verify their equality. As many readers know, verifying things that are supposedly obvious often leads to
the discovery of problems.
The claim that the above two tensorizations are the same is not general in the abstract. It is specific in
that it applies in the "tensor world" built upon a transformation x' = F(x) where x-space is Cartesian. The
two forms are not equal if x-space has some metric tensor other than g = 1.

Our task then is to show that the following is true when x-space is Cartesian,

214
 Chapter 15: Covariant Method

B'n;j;j = B'j;j;n – g'-1/2ε'nab(g'-1/2g'bcε'cdeB'e;d);a ? (15.8.3)

The plan is to rewrite the above equation in a sequence of reducing steps until we get to an equation that
obviously is true. Then one then can reverse the sequence to conclude that (15.8.3) is true. We shall mark
each step with a question mark to the right, to indicate that this is "something we want to show is true".
When the final step is reached which is true, the reader then goes backwards and erases the question
marks.

As a first and rather large step in terms of clutter reduction, we claim that,

(g'-1/2g'bcε'cdeB'e;d);a = g'-1/2g'bcε'cde (B'e;d);a . (15.8.4)

This is an application of several facts derived in Appendix F. The main idea is Theorem (F.10.15)
concerning what objects are "extractable" from a covariant derivative expression. Eq. (F.10.17) shows
that ε'cde is extractable since ε'cdeα; = 0, while (F.10.18) shows that g'bc is extractable since g'bc;α= 0.
Finally, (F.10.20) says that, for any real s, g's is extractable since (g's);α = 0, recalling that g' > 0. We
then arrive at

B'n;j;j = B'j;j;n – g'-1/2ε'nab(g'-1/2g'bcε'cdeB'e;d);a

= B'j;j;n – g'-1/2ε'nabg'-1/2g'bcε'cde (B'e;d);a // now use g'bc and note ε'nab = ε'ban

= B'j;j;n – g'-1ε'ban ε'bde B'e;d;a .

Therefore

B'n;j;j = B'j;j;n – g'-1ε'bdeε'ban B'e;d;a ? (15.8.5)

We would next like to have all the B' tensors be multiples of B'e;d;a which appears in the last term, so :

B'n;j;j = B'n;a;a = g'ne g'adB'e;d;a

B'j;j;n = B'd;d;n = g'de g'na B'e;d;a . (15.8.6)

Installing these expressions into (15.8.5) gives

g'ne g'adB'e;d;a = g'de g'na B'e;d;a – g'-1 ε'bdeε'bna B'e;d;a ?

or
[ g'ne g'ad - g'de g'na + g'-1 ε'bde ε'bna ] B'e;d;a = 0 ? (15.8.7)

At this point, one might hope that [....] = 0 and we are done. If this were true, our claim (15.8.3) would be
true for any metric tensor g, but we know that is not the case, so we do not expect to have [...] = 0. The
next step is to call upon (D.11.11) expressed in x'-space which says

215
 Chapter 15: Covariant Method

g'-1 ε'sABε'sA'B' = g'AA'g'BB' – g'BA'g'AB' (D.11.11)

or
g'-1 ε'bdeε'bna = g'dng'ea – g'en g'da . (15.8.8)

Putting this into (15.8.7) gives

[ g'ne g'ad - g'de g'na + g'dng'ea – g'eng'da ] B'e;d;a = 0 ?

or
[ - g'de g'na + g'dng'ea ] B'e;d;a = 0 ?
or
[ g'dng'ea - g'deg'na ] B'e;d;a = 0 ? (15.8.9)

As expected, we find [...] ≠ 0. We have to use the fact that x-space is Cartesian. To this end write the
covariant transformation rule for rank-3 tensor Be;d;a as shown for example in (7.10.1),

B'e;d;a = ReERdDRaA BE;D;A = ReERdDRaA BE,D,A = ReERdDRaA ∂D∂A BE (15.8.10)

where the critical Cartesian x-space fact will be that ∂D∂A BA is symmetric under A↔D , Inserting the
above transformation rule into (15.8.9) gives,

[ g'dng'ea - g'deg'na ] ReERdDRaA∂D∂A BE = 0 ?

or
g'dng'ea ReERdDRaA ∂D∂A BE = g'deg'na ReERdDRaA ∂D∂A BE ? (15.8.11)

At this point we recall from (7.5.9) that g'** raises the first index of R** , whereas the second index of
R** goes up and down for free since g = 1 in x-space, so we will take second indices all down. Thus,

RaERnDRaA∂D∂A BE = RdERdDRnA ∂D∂A BE ?

or
RaERnDRaA∂D∂A BE = RdERdDRnA∂D∂A BE ? (15.8.12)

Now on the right side do D↔A on these dummy indices, but then restore the ∂D∂A order so

RaERnDRaA∂D∂A BE = RdERdARnD∂D∂A BE ? (15.8.13)

and changing d→a on the right

RaERaARnD ∂D∂A BE = RaERaARnD∂D∂A BE ? (15.8.14)

The two sides are exactly the same, so we are done and thus (15.8.3) is verified.

216
 Appendix A: Reciprocal Base Vectors

Appendix A: Reciprocal Base Vectors the Hard Way

A.1 Introduction

Note: This Appendix is written in the developmental notation, not the Standard Notation, though a few
equations are translated to the latter form. The rules for translation to Standard Notation are

(7.13.1) (7.5.2) (7.5.4) (7.4.1) (7.4.1)

En → en Rij → Rij Sij → Sij ḡ'nm → g'nm g'nm → g'nm .

The tangent base vectors en are defined in Chapter 3, while the reciprocal base vectors En are defined in
Chapter 6. Here are some facts regarding these vectors gathered from those chapters:

(3.2.5) and (6.1.4) (6.2.4) (6.2.4) and (6.1.2) (3.2.7) and (6.6.3) and (6.1.2)
(en)k = Skn en • em = ḡ'nm |en| = ḡ'nn = h'n S = [e1, e2, e3 .... eN ]
(En)i ≡ gia Rna En • Em = g'nm |En| = g'nn R = [Ē1, Ē2, Ē3 .... ĒN ]T
= g'na Sia en • Em = δn,m En ≡ g'ni ei en = ḡ'ni Ei . (A.1.1)

These facts are all applicable in the Picture A context with arbitrary metric tensors g' and g,

(A.1.2)

This Appendix begins with a different definition of something called Ek. Although the definition is
meaningful in the general Picture A context, the object so defined only agrees with the Ek of Chapter 6 if
x-space is Cartesian (g = 1). The reason can be traced to the fact that the dot product rule en • Em = δn,m
is only valid for the Appendix A definition of Em when g = 1 because only then is a cross product
orthogonal to all its component vectors. One application of the reciprocal base vectors is in the study of
curvilinear coordinates where one always takes g = 1, and g' is then the curvilinear coordinates metric
tensor of interest. Therefore, the reader should think of this Appendix in the context of Picture B

(A.1.3)

217
 Appendix A: Reciprocal Base Vectors

A.2 Definition of En

The reciprocal base vectors are defined in the following very strange looking and clumsy manner,

(Ek)α ≡ det(R) (-1)k-1 εαi1i2i3...ik...iN (e1)i1 (e2)i2 ...... (ek)ik.......... (eN)iN (A.2.1)

where N is the number of dimensions of the Cartesian x-space RN in which the vectors en and En exist.
Notice that the ε subscript ik is "crossed out" and the same for factor (ek)ik . Crossed out means they are
simply missing, they are omitted. Thus, in the above expression there are N-1 implied summation indices
(α is fixed) and there are N-1 factors of the form (en)in .
The object ε has N subscripts and is the "totally antisymmetric tensor" in N dimensions: ε123...N ≡
+1, and each time any two indices on ε are swapped, ε negates. For example, ε1234 = 1 but ε1432 = -1. If
two indices are the same, then ε = 0.

A.3 Simpler notation

To avoid dealing with subscripts on subscripts, one can rewrite the above definition in a less precise but
simpler notation

(Ek)α ≡ det(R) (-1)k-1εαabc...x (e1)a(e2)b ...... (eN)x // κ(k) and (ek)κ are missing (A.3.1)

In this notation, subscript x stands for the Nth letter of the alphabet (imagine N ≤ 26). If κ is the kth letter
of the alphabet, then κ is missing from the indices on ε, and the factor (ek)κ is missing from the product of
factors. For example, if k = 2, then summation index κ = b is missing from the ε.
Now take the ε subscript α and slide it right to the "hole" where κ is missing, picking up a minus sign
for each step of this slide. Moving k-1 positions results in (-1)k-1. Thus the above becomes,

(Ek)α ≡ det(R) εabc..α..x (e1)a(e2)b ...... (eN)x // (ek)κ is missing, α in κ position (the kth) (A.3.2)

Example: For N = 3 the above becomes,

(E1)α ≡ det(R) εαbc(e2)b(e3)c ⇒ E1 = det(R) e2 x e3 a is missing

(E2)α ≡ det(R) εaαc(e1)a(e3)c ⇒ E2 = det(R) e3 x e1 b is missing
(E3)α ≡ det(R) εabα(e1)a(e2)b ⇒ E3 = det(R) e1 x e2 c is missing (A.3.3)

and the results are cyclic. Here is a detail from the middle line

εaαc(e1)a(e3)c = – εαac(e1)a(e3)c = + εαca(e1)a(e3)c = εαca(e3)c (e1)a = [ e3 x e1 ]α (A.3.4)

A.4 Generalized Cross Product of N-1 vectors of dimension N

One can define a generalized "cross product" of N-1 vectors, each of dimension N, in this fashion:

Qa ≡ εabc...x BbCcDd.....Xx (A.4.1)

218
 Appendix A: Reciprocal Base Vectors

where x and X represent the Nth letter of the alphabet. The ε object is again the totally antisymmetric
tensor with N indices. In vector notation one writes this symbolically as

Q = B x C x D x ... x X / N-1 factors, N-2 crosses . (A.4.2)

This vector notation is defined by the previous line.

The vector Q is orthogonal to all the vectors from which it is constructed! For example (here is the
point where Q • C ≡ gabQaCb needs to be QaCa, so g = 1 is required in x-space)

Q • C = CaQa = Ca εabc...x BbCcDd.....Xx = BbDd...Xx { εabc...x CaCc } . (A.4.3)

But {..} is the contraction of something symmetric under a↔c (CaCc) with something antisymmetric
under a↔c (εαabc...x) and therefore {..} = 0. In general,

SacAac = Sca Aca // relabel both dummy summation indices

= Sac (-Aac) // S is Symmetric, A is antisymmetric
= - Sac Aac // = the negative of the starting expression
=0 . (A.4.4)

Similarly, Q•A = 0, Q•B = 0 and so on.

Swapping the position of any two vectors in the generalized cross product causes Q to change sign.
For example, swapping B and C,

Qa ≡ εabc...x CbBcDd.....Xx = εacb...x CcBbDd.....Xx // b ↔ c

= - εabc...x BbCcDd.....Xx = -Qa . // swap indices on ε (A.4.5)

Thus, the notions of orthogonality and interchange are consistent with the regular Q = B x C cross
product for N=3.
When N=2, one must be a little careful with this notation. The component equation is

Qa ≡ εab Bb ⇒ Q1 = B2 and Q2 = -B1 . (A.4.6)

One might be tempted to express the vector equation as Q = B since there are no "no crosses". This
vector equation is wrong, while the component equation is correct. One can rescue the vector notation by
a simple trick. When N=2 the vector B can be represented of course as B = B ^1+B ^
1 2 . Imagine this 2D
2

space to be embedded in the usual 3D space with a third axis ^

3 . Then consider this 3D cross product:

Q=Bx^
3 ⇒ Qa ≡ εabc Bb(^
3 )c = εabc Bbδ3,c = εab3 Bb = εabBb . (A.4.7)

Thus, this trick reproduces the correct component equation, and it makes more obvious the fact that Q is
orthogonal to B.
Summary: The generalized cross product Q of N-1 vectors each of dimension N can be expressed in
both component and vector notation:

219
 Appendix A: Reciprocal Base Vectors

Qa ≡ εabc...x BbCcDd.....Xx
Q = B x C x D x ... x X / N-1 factors, N-2 crosses (A.4.8)

Q is orthogonal to all the vectors from which it is composed. Swapping any two vectors negates Q. When
N=2, one can rescue the otherwise failing vector notation by thinking of it as saying Q = B x ^
3.

Comment: Notice that Q = B x C x D is defined for 4-vectors only. This is a completely different animal
from the object Q = B x (C x D) which is defined for 3-vectors only. This latter object contains two ε
factors, while the former only one.

A.5 Missing Man Formation

We now make a small variation in the notation. Start with the above equation,

Qa ≡ εabc...x BbCcDd.....Xx , (A.5.1)

then change a to α, back up all the Latin letters by one (but leave the last as "unknown" x), and assume
that some subscript κ and factor Kκ are "missing". The result is,

Qα ≡ εαac...x AaBbCc.....Xx . // κ and Kκ are missing (A.5.2)

There are still N-1 factors, and one can still write this in vector notation

Q = A x B x C x ... x X // K is missing (A.5.3)

and of course it is still true that Q•C = 0, etc. For N=2 the vector notation is rescued as in (A.4.7) above.

A.6 Apply this Notation to E

Compare the above Qα of (A.5.2) to the {...} part of the (A.3.1) definition of (Ek)α ,

(Ek)α ≡ det(R) (-1)k-1{ εαabc...x (e1)a(e2)b ...... (eN)x } // κ(k) and (ek)κ are missing; N≥ 2
(A.3.1) (A.6.1)
Therefore, the definition of Ek for N > 2 can be written in this vector notation,

Ek ≡ det(R) (-1)k-1 e1 x e2 x ......x eN // ek missing; N > 2 . (A.6.2)

The reciprocal base vector Ek is thus orthogonal to all the tangent base vectors from which it is
constructed (remember ek is missing)! For example, for N=3 the three E vectors are given by

E1 = det(R) (-1)1-1 e2 x e3 = det(R) e2 x e3

E2 = det(R) (-1)2-1 e1 x e3 = det(R) e3 x e1
E3 = det(R) (-1)3-1 e1 x e2 = det(R) e1 x e2 (A.6.3)

220
 Appendix A: Reciprocal Base Vectors

which agrees with the results (A.3.3). For N =2 ( E's label corresponds to the missing e's label ),

E1 = det(R) (-1)1-1 e2 x ^
3 = det(R) e2 x ^ 3 or (E1)k = det(R) εka(e2)a
E2 = det(R) (-1)2-1 e1 x ^
3 = - det(R) e1 x ^
3 or (E2)k = -det(R) εka(e1)a (A.6.4)

One can combine these two lines into one as follows ( eg, k = 1, then 3-1 = 2, etc)

Ek = det(R) (-1)k-1 e3-k x ^

3 = det(R) e3-k x ^
3 or (E1)k = det(R) (-1)k-1εka(e3-k)a . (A.6.5)

The vector "trick" notation shows that E1•e2 = 0 and E2•e1 = 0,

E1•e2 = det(R) e2 x ^
3 • e2 = 0
E2•e1 = -det(R) e1 x ^
3 • e1 = 0 (A.6.6)
and also
E1•e1 = det(R) εka(e2)a (e1)k = det(R)det[e1, e2] = det(R)det(S) = 1
E2•e2 = -det(R) εka(e1)a (e2)k = -det(R)det[e2, e1] = det(R)det(S) = 1 . (A.6.7)

It is shown next that these N=2 results are special cases of a general fact: Em • en = δm,n .

Eq. (5.11.5) showed that em • en = ḡ'mn . The other two dot products are now considered.

A.7 Compute Em • en

One can now compute, for general N,

Ek • ek = (Ek)α(ek)α = { det(R) (-1)k-1εαabc...x (e1)a(e2)b ...... (eN)x } (ek)α . (A.7.1)

k is missing

Slide α to the right in the ε subscript field and put it into the hole of the missing subscript κ, picking up
(-1)k-1. At the same time, move the (eκ)α to the left and position it in its proper place in the product of
factors,

Ek • ek = (Ek)α(ek)α = { det(R) εabc..α..x (e1)a(e2)b ... (eκ)α ... (eN)x }

= det(R) det [e1, e2, e3 .... eN ] = det(R) det(S) = det(RS) = det(1) = 1 (A.7.2)

where S = [e1, e2, e3 .... eN ] from (3.2.7). We already know that Ek is orthogonal to all the en which form
the generalized cross product, therefore

Em • en = δm,n (A.7.3)

which is the "duality relation" discussed more generally in Section 6.2.

221
 Appendix A: Reciprocal Base Vectors

A.8 Compute En • Em

Since the vectors { en } are linearly independent and thus form a basis in RN, Em can be expanded onto
the en ,

Em = Σn An(m) en (A.8.1)
δm,k = Em • ek = Σn An(m) en • ek = Σn An(m) ḡ'nk . (A.8.2)

Multiplying both sides of (A.8.2) by g'ki and summing on k gives

LHS = Σk g'ki δm,k = g'mi

RHS = Σn An(m) (Σk ḡ'nk g'ki) = Σn An(m) (ḡ'g')ni = Σn An(m)δn,i = Ai(m) . (A.8.3)

Therefore Ai(m) = g'mi so,

Em = Σn An(m) en = Σn g'mn en (A.8.4)

which is to say Ek is this linear combination of the ei, which is the definition used in Chapter 6, (6.1.2),

Ek = Σi g'ki ei = g'ki ei // implied sum on i // Std Notation: ek = Σi g'ki ei (A.8.5)

This may be compared with the previous result (A.6.2),

Ek ≡ det(R) (-1)k-1 e1 x e2 x ......x eN // ek missing; (A.6.2)

It seems rather impressive that these two dissimilar ways of writing E are equal. Finally,

En • Em = En • (g'mi ei) = g'mi (En • ei) = g'mi δn,i = g'mn = g'nm // g' is symmetric (A.8.6)

A.9 Summary of relationship between the tangent and reciprocal base vectors

en • em = ḡ'nm En • Em = g'nm en • Em = δn,m

En = Σi g'ni ei en = Σi ḡ'ni Ei ḡ' = g'-1 . (A.9.1)

Although these results have just been derived in the Picture B context, they are also valid in the more
general Picture A context, as shown in Chapter 6 in which the equation En = Σi g'ni ei is used as the
definition of En . As a reminder, the cross product expression for En is only valid in Picture B where g=1.

In Standard Notation, the summary above can be restated as

en • em = g'nm en • em = g'nm en • em = δnm

en = Σi g'ni ei en = Σi g'ni ei g'ab = (g'ab) -1 (A.9.2)

222
 Appendix A: Reciprocal Base Vectors

A.10 Another Cross Product Notation and another expression for E

Go back to the general cross product of N-1 vectors each of dimension N,

Q = B x C x D x ... x X . // N-1 factors, N-2 crosses (A.4.2)

Replace B,C,D ... by vectors A(n),

Q = A(1) x A(2) x A(3) x ... x A(N-1) // N-1 factors, N-2 crosses (A.10.1)

It is convenient to write this using a product symbol Πx ,

Q = Πxi=1N-1 A(i) = Πxi A(i) (A.10.2)

where in the second form it is understood that i takes on all values i = 1 to N-1. The superscript x means
that this is not a regular product, it is our generalized cross product. This Πx symbol also implies correct
handling of the special case N=2 such that

Q = Πxi=11 A(i) ≡ A(1) x ^

3 // ≠ A(1) (A.10.3)

as discussed in (A.4.7) above.

This same Πx notation can be applied to the "missing man formation" of Section A.5 above. Suppose

Q = A(1) x A(2) x A(3) x ... x A(N) // A(n) is missing . (A.10.4)

One can write this as

Q = Πxi=1..N,i≠n A(i) ≡ Πxi≠n A(i) . (A.10.5)

And of course this idea can be applied to the expression for Ek

Ek = det(R) (-1)k-1 e1 x e2 x ......x eN // ek missing; (A.6.2)

Ek = det(R) (-1)k-1 Πxi≠k ei . (A.10.6)

Once again, for N=2 the Πx symbol implies that ( ek = "missing", e3-k = the one not missing)

Πxi≠k ei = Πxi=1..2,i≠k ei = e3-k x ^

3 (A.10.7)

Ek = det(R) (-1)k-1 e3-k x ^

3 (A.10.8)

which is the "trick" notation of s(A.4.7) above for the N=2 case.

223
 Appendix B: Geometry of N-pipeds

Appendix B: The Geometry of Parallelepipeds in N dimensions

B.1 Overview

This Appendix presents a simple method for constructing an N dimensional parallelepiped, which name
we shorten to "N-piped". It is found that an N-piped has 2N vertices and N pairs of faces for a total of 2N
faces, and the locus of points that make up each of these faces is stated. Each face of an N-piped is in fact
an (N-1)-piped which has 2N-1 vertices and is planar in N dimensions (meaning it lies on an N-1
dimensional flat surface). The two faces which make up each face pair lie on parallel planes in RN.
For example, for N=3 each face is a 2-piped having 23-1 = 4 vertices, and there are N=3 face pairs for
a total of 6 faces, and each pair of faces is planar in 3 dimensions.
For N=4 there are 4 pairs of faces for a total of 8 faces. Each face is a 3-piped having 24-1 = 8
vertices. For example, one would say that each face of a 4-cube is a 3-cube. It is not intuitively obvious
that two faces each of which is a regular cube can in fact lie on surfaces which are planar and parallel in 4
dimensions, but we show how this works below.
It is then shown that, if the N-piped is spanned by the N tangent base vectors en of Chapter 3, the
normal vectors for the pairs of parallel faces are just the reciprocal base vectors En of Chapter 6.
Section B.5 focuses on the area and volume of N-pipeds in various dimensions, and simple
expressions for the volume and vector areas of the faces of an N-piped are obtained.
Rather than just state the results in N dimensions, we attempt an inductive approach to provide
motivation for the N dimensional results. In this approach, cases N = 2,3.. are treated with nearly identical
boilerplate templates to build up the inductive case.
A reader interested in the details here should read Appendix A first since use is made of various
Appendix A results including the generalized cross product idea. On the other hand, a reader not
interested in details might just read the Appendix B Summary presented below in Section B.6.

B.2 Preliminary: Equation of a plane in N dimensions

Consider an arbitrary plane drawn in N space which does not pass through the origin. There is some point
on that plane which lies closer to the origin than all other points on the plane. Let p be a vector from the
origin to that closest point, and let r represent a point lying on the plane,

(B.2.1)

Since p is normal to the plane, and since r-p is a vector lying in the plane, it follows that

p•(r-p) = 0 ⇒ r•p = p2 ⇒ r•^

p=p . (B.2.2)

Therefore, one way to write the equation of a plane in N dimensions is

224
 Appendix B: Geometry of N-pipeds

r•^
p=p r = (x1, x2, .....xN) . (B.2.3)

where ^p is the unit vector normal to the plane which points "away from the origin", and where p > 0 is the
distance of closest approach of the plane to the origin. In the limit p→0, the plane passes through the
origin and the equation is then r•^
p = 0 where ^p is either normal to the plane.

B.3 N-pipeds and their Faces in Various Dimensions

(a) The 1-piped

Start with N = 1 where the piped is some arbitrary line segment e1 in direction ^
e 1 having length e1, with
one end affixed to the origin of the real axis :

(B.3.a.1)

This piped has two vertices located at v1 = 0 and v2 = e1. These two vertices are also the "faces" of this
1-piped, so there are two faces (one pair of faces). These faces are 0 dimensional and therefore don't point
in any direction (they are the endpoints of the vector e1). The 1-piped is a piece of a plane in 1 dimension
(a line). One can think of the vertex at the origin as the "generator 0-piped" and the other vertex as the
partner face of the generator, in the sense of the generator idea described below.

The loci of points in the 1-piped's interior "volume" is given by

rvolume1 = α1 e1 0 ≤ α1 ≤ 1 . (B.3.a.2)

The volume of this 1-piped is e1.

(b) The 2-piped

Now add another dimension, going to N=2. Introduce a unit vector ^ e 2 in some arbitrary direction in R2
other than e1 so that e1 and e2 are linearly independent. Take the 1-piped described above (line segment)
and translate it by e2 to create a new copy of the line segment. The original 1-piped we call the generator
piped, and the copy is the partner of the generator piped which, it will be shown, lies on a plane (a 1-plane
= line) which is parallel to the plane of the generator piped, but its plane does not pass through the origin.
In N=2 dimensions, the generator 1-piped and its partner are now "faces" of a 2-dimensional object, a
parallelogram = a 2-piped. Draw line segments from all the vertices of the generator piped to matching
vertices of its partner piped (add 2 line segments) to make 2 additional "side" faces. One of these faces
necessarily touches the origin, and the other face does not. Faces always occur in parallel pairs, one face
of which touches the origin, and one of which does not, the latter we will call the "partner" face. For our
2-piped, each face is a 1-piped. There are now four faces, each is a line segment.

225
 Appendix B: Geometry of N-pipeds

(B.3.b.1)

The loci of the 2-piped's volume and of its four 1-piped faces are given by

rvolume2 = α1e1 + α2 e2 0 ≤ α1,α2 ≤ 1

rface2 = α1e1 0 ≤ α1 ≤ 1 // the generator face

rface2p = α1e1 + e2 0 ≤ α1 ≤ 1 // partner of the generator face
rface1 = α2e2 0 ≤ α2 ≤ 1 // side face touching the origin
rface1p = α2e2 + e1 0 ≤ α2 ≤ 1 // partner of the above side face (B.3.b.2)

The origin-touching faces are numbered using the index of the en vector that does not appear in the locus
for the face. This seems strange but for N > 2 it will be clear why this is done. The above results can be
summarized as

rvolume3 = Σnαnen 0 ≤ αn ≤ 1

rface(i) = Σn≠iαnen 0 ≤ αn ≤ 1 i = 1,2,3

rface(ip) = Σn≠iαne + ei 0 ≤ αn ≤ 1 i = 1,2,3 (B.3.b.3)

According to Note 1 following (6.2.7), it is possible to construct vectors E1 and E2 as linear combinations
of e1 and e2 such that the following is true:

Ei• ej = δi,j // Ek = Σi=12 g'ki ei, see Note 4 following (6.2.7) (B.3.b.4)

If one interprets the en vectors as tangent base vectors for some transformation F, then the two vectors En
are the corresponding reciprocal base vectors which are discussed in Chapter 6 and Appendix A.

Consider now these dot products:

E2 • rface2 = E2 • α1e1 = 0 ⇒ ^ •r
E2 face2 = 0

E2 • rface2p = E2 • [ α1e1+ e2] = 1 ⇒ ^ •r

E2 face2p = 1/E2 . (B.3.b.5)

According to (B.2.3), the first line says that face 2 lies on a plane which passes through the origin and
which has normal vector E ^ . Also according to (B.2.3), the second line says that face 2p has the same
2
normal and its plane is therefore parallel to face 1 but misses the origin by distance 1/|E2|. Similarly,

E1 • rface1 = E1 • α2e2 = 0 ⇒ ^ •r
E1 face1 = 0

E1 • rface1p = E1 • [ α2e2+ e1] = 1 ⇒ ^ •r

E1 face1p = 1/E1 . (B.3.b.6)

226
 Appendix B: Geometry of N-pipeds

^ . The first touches the origin while the partner's

These two faces are also parallel, both having normal E1
plane misses the origin by distance 1/|E1| .

The conclusions that En is normal to face n, and that the pair of faces n and np are parallel, do not depend
on the specific upper endpoints of the ranges of α1 and α2 which happen to be given as 1 above. This
seems pretty obvious since rescaling the edges of a parallelogram does not affect its normal vector.

(c) The 3-piped

Now add another dimension, going to N=3. Introduce a unit vector ^ e 3 in some arbitrary direction in R3 so
that (^
e 1,^
e 2,^
e 3) are linearly independent. Take the 2-piped described above (parallelogram) and translate
it by distance e3 in the ^
e 3 direction to create a new copy of the 2-piped. The original 2-piped we call the
generator piped, and the copy is the partner of the generator piped which, as will now be shown, lies on a
plane which is parallel to that of the generator piped, but which does not pass through the origin. In N=3
dimensions, the generator 2-piped and its partner are now "faces" of a 3-dimensional object, a
parallelepiped = a 3-piped. Draw line segments from all 22 vertices of the generator piped to the
corresponding vertices of its partner piped (add 4 line segments), to get 4 additional side faces. Two of
these faces necessarily touch the origin, and the other two do not. For the 3-piped, each face is a 2-piped.
There are now 2*3 = 6 faces, each is a 2-piped.

(B.3.c.1)

The loci of the 3-piped's volume and of its six 2-piped faces are given by

rvolume3 = α1e1 + α2 e2 + α3 e3 0 ≤ α1,α2,α3 ≤ 1

rface3 = α1e1 + α2 e2 0 ≤ α1,α2 ≤ 1 // the generator face

rface3p = α1e1 + α2 e2 + e3 0 ≤ α1,α2 ≤ 1 // partner face to the above
rface2 = α1e1 + α3 e3 0 ≤ α1,α3 ≤ 1 // the generator face
rface2p = α1e1 + α3 e3 + e2 0 ≤ α1,α3 ≤ 1 // partner face to the above
rface1 = α2e2 + α3 e3 0 ≤ α2,α3 ≤ 1 // the generator face
rface1p = α2e2 + α3 e3 + e1 0 ≤ α2,α3 ≤ 1 // partner face to the above (B.3.c.2)

Notice that the partner face locus is created from the non-partner face by adding "the other" base vector.
For example, face 3 is "spanned" by base vectors e1 and e2 so e3 is added to get the partner face face3p. A

227
 Appendix B: Geometry of N-pipeds

partner is just a copy of the non-partner which is translated by a constant vector. The above results can be
summarized in this concise manner:

rvolume3 = Σnαnen 0 ≤ αn ≤ 1

rface(i) = Σn≠iαnen 0 ≤ αn ≤ 1 i = 1,2,3

rface(ip) = Σn≠iαne + ei 0 ≤ αn ≤ 1 i = 1,2,3 (B.3.c.3)

According to Note 1 following (6.2.7), it is possible to construct vectors E1, E2, E3 as linear
combinations of e1, e2, e3 such that the following is true,

Ei• ej = δi,j // Ek = Σi=13 g'ki ei, see Note 4 following (6.2.7) (B.3.c.4)

If one interprets the en vectors as tangent base vectors, then the three vectors En are the corresponding
reciprocal base vectors which are discussed in Chapter 6 and Appendix A. Consider now these dot
products:

E1• rface1 = E1• [ α2e2 + α3 e3] = 0 ⇒ ^ •r

E1 face1 = 0

E1• rface1p = E1• [α2e2 + α3 e3 + e1] = 1 ⇒ ^ •r

E1 face1p = 1/E1 . (B.3.c.5)

According to (B.2.3), the first line says that face 1 lies on a plane which passes through the origin and
which has normal vector E ^ . Also according to (B.2.3), the second line says that face 1p has the same
1
normal and its plane is therefore parallel to face 1 but misses the origin by distance 1/|E1| .
A similar pair of equations obtains for each of the other face pairs.

(d) The N-piped

Now add another dimension, going from N-1 to N. Introduce a unit vector ^ e N in some arbitrary direction
in R so that ( ^
N
e 1... ^
e N ) are linearly independent. Take the (N-1)-piped described above and translate it
by distance eN in the ^ e N direction to create a new copy of the (N-1)-piped. The original (N-1)-piped we
call the generator piped, and the copy is the partner of the generator piped which, it will be shown, lies on
a plane which is parallel to that of the generator piped, but which does not pass through the origin. The
generator (N-1)-piped and its partner are now "faces" of a N-dimensional object, an N-piped. Adding this
partner piped doubles the total vertex count. Draw line segments from all 2N-1 vertices of the generator
piped to the corresponding vertices of its partner piped to get 2N-2 additional side faces for a total now of
2N faces. There are N pairs of "faces" because there are N ways to omit a single ei from the list of vectors
which span a face, so including the partner faces an N-piped has 2N faces in total. Half of these faces
necessarily touch the origin, and the other half do not. Each face is an (N-1)-piped.
It is convenient to refer to the partner face of a pair as "the far face" and the other one, which touches
the origin, as "the near face".
We leave the N-piped drawing which would be labeled (B.3.d.1) to the reader's imagination, and we
omit a detailed list of the face volumes which we could called (B.3.d.2), giving just the summary below.

228
 Appendix B: Geometry of N-pipeds

The loci of the N-piped's volume and of its 2N (N-1)-piped faces are given by:

rvolumeN = Σnαnen 0 ≤ αn ≤ 1
rface(i) = Σn≠iαnen 0 ≤ αn ≤ 1 i = 1,2...N
rface(ip) = Σn≠iαnen + ei 0 ≤ αn ≤ 1 i = 1,2...N (B.3.d.3)

Notice that the partner face locus is created from the non-partner face by adding "the other" base vector.
For example, face i is "spanned" by base vectors en n ≠ i, so it is ei that one adds to get the partner. A
partner is just a copy of the non-partner which is translated by a constant vector.

According to Note 1 following (6.2.7), it is possible to construct vectors E1...EN as linear combinations
of e1.. eN such that the following is true,

Ei• ej = δi,j // Ek = Σi=12 g'ki ei, see Note 4 following (6.2.7) (B.3.d.4)

If one interprets the en vectors as tangent base vectors, then the N vectors En are the corresponding
reciprocal base vectors which are discussed in Chapter 6 and Appendix A. Consider now these dot
products:

Ei• rface(i) = Ei• [Σn≠iαnen] = 0 ⇒ ^ •r

Ei face(i) = 0

Ei• rface(ip) = Ei• [Σn≠iαne + ei] = 1 ⇒ ^ •r

E = 1/E (B.3.d.5)
i face(ip) i

The first line says that face i lies on a plane which passes through the origin and which has normal vector
^ . The second line says that face ip has the same normal and its plane is therefore parallel to face i but
Ei
misses the origin by distance 1/|Ei| .

B.4 The question of inward versus outward facing normal vectors

It has been shown above that, for an N-piped, the pair of faces i and ip has normal vector Ei. For one of
these faces, Ei will be an outward directed normal, while for the other it will be an inward directed
normal. One might like to know which is which. Here is one way to find out.

First, construct these three vectors

(piped)center = Σn(1/2)en // vector from origin to piped center

(face i)center = Σn≠i(1/2)en // vector from origin to center of face i

(face ip)center = Σn≠i(1/2) en + ei // vector from origin to center of face ip . (B.4.1)

Construct vectors from piped center to face centers ( results here are fairly obvious)

229
 Appendix B: Geometry of N-pipeds

(face i)center - (piped)center = [Σn≠i(1/2)en] - Σn(1/2)en = - (1/2)ei

(face ip)center - (piped)center = [Σn≠i(1/2)en+ ei ] - Σn(1/2)en = + (1/2)ei . (B.4.2)

Then compute

Ei • {(face i)center - (piped)center } = Ei • [- (1/2)ei ] = -(1/2) < 0

Ei • {(face ip)center - (piped)center } = Ei • [+ (1/2)ei ] = +(1/2) > 0 . (B.4.3)

One may conclude that Ei is an outward pointing normal for face ip (the far face of the pair). Therefore,
- Ei is an outward pointing normal for face i (the near face of the pair), which recall is the face which
touches the origin.. Here is an illustration for the case N = 2:

(B.4.4)

B.5 The Face Area and Volume of N-pipeds in Various Dimensions

We embark now on another long march to inductively arrive at results for the general N case. Tracing the
first few cases N = 2,3,4 and then extrapolating to N = N probably gives more insight than a formal
induction proof which is not attempted here. Each case below is treated with the same boilerplate
template which first treats Face Area and then Volume.

(a) The 2-piped

For this case N = 2, the reader is encouraged to stare at the above Fig (B.4.4) while reading the text.

Face Area. The area of a 2-piped face (a line segment) is just the length of the edge which is the face,

A1 = |e2|
A2 = |e1| . (B.5.a.1)

where here we maintain the plan of labeling an area by the index of the spanning vector which is omitted
in making the area. The vector areas can be written, based on the work above,

^
A1 = |e2| E // Ek = Σi g'ki ei , see Note 4 following (6.2.7)
1

A = |e | E^ (B.5.a.2)
2 1 2

and these vectors are out-facing for faces 1p and 2p. We claim that both these results can be expressed in
a single formula,

230
 Appendix B: Geometry of N-pipeds

An = |det(S)| En . (B.5.a.3)

One can see that the direction is correct for n = 1,2, so it is just a question of verifying the magnitude. One
must show that

|det(S)| |E1| = |e2| and |det(S)| |E2| = |e1| (B.5.a.4)

or
|Ek| = |det(R)| |e3-k| k=1,2 . // RS = 1 (B.5.a.5)

Using the N=2 trick notation from (A.6.5),

Ek = det(R) (-1)k-1 e3-k x ^

3 (A.6.5) (B.5.a.6)
so that
| Ek | = | det(R)| | e3-k x ^
3 | = |det(R)| | e3-k| k = 1,2 (B.5.a.7)

since e3-k and ^

3 are perpendicular. QED.
We stress the formula An = |det(S)| En because it will turn out that this is valid for all N ≥ 2 .
One can restate An = |det(S)| En using the cross product notation presented in (A.10.6):

An = |det(S)| En = |det(S)| det(R) (-1)n-1 Πxi≠n ei

= σ (-1)n-1 Πxi≠n ei σ ≡ sign(det(S)) = sign(det(R)) (B.5.a.8)

with the understanding that for the N=2 case Πxi≠n ei has the special meaning of (A.10.7), so that in fact

An = σ (-1)n-1 [ Πxi≠n ei ] = σ (-1)n-1 [ e3-n x ^

3] // for N = 2, n = 1,2 (B.5.a.9)

Volume. The volume of a 2-piped is the base times the height of a parallelogram, familiarly given by the
cross product of the edges,

volume(2) = | e1 x e2 | = | εab (e1)a(e2)b | = | det [ e1, e2 ] | = | det(S) | // see (3.2.7) (B.5.a.10)

where S is the linearized transformation matrix for N=2, see Chapter 2. Of course strictly in N=2 the
notation e x e has no meaning, so one has to imagine a ^
1 2 3 dimension to give it meaning. The second
form does have a meaning for N=2, and that meaning is |(e1)1 (e2)2 – (e1)2 (e2)1|.

(b) The 3-piped

Face Area: The faces of a 3-piped are 2-pipeds. For N=2, the 2-piped volume is this from (B.5.a.9),

volume(2) = | εab (e1)a(e2)b | , (B.5.a.9) (B.5.b.1)

231
 Appendix B: Geometry of N-pipeds

where e1 and e2 are 2D vectors. For the 2-piped which is "face 3" of the 3-piped -- a "near" face which
touches the origin of the 3D skewed en coordinate system -- vectors e1 and e2 are 3D vectors. This face 3
is the top face shown in Fig (B.3.c.1) above. The first 2 components of each of these 3D vectors are the
same as the components of the 2D ei vectors, while the 3rd components are both 0. This is so because
face 3 lies in a plane defined by this 3rd component being 0. The volume(2) formula expressed in terms of
these new 3D vectors is therefore |εab3 (e1)a(e2)b|, where ε is now a 3D ε tensor. The conclusion is that
the area of the top face in Fig (B.3.c.1) is,

A3 = |εab3 (e1)a(e2)b| (B.5.b.2)

and this then is the scalar area of both face 3 and its partner face 3p, the far face. Similar arguments would
then support these other area expressions

A1 = |ε1ab (e2)a(e3)b|
A2 = |εa2b (e3)a(e1)b| . (B.5.b.3)

Since indices on ε can be swapped for free due to the absolute value, the non-summed index can be put
first in all three cases and one may then conclude that

A1 = |e2 x e3| face 1 and face 1p

A2 = |e3 x e1| face 2 and face 2p
A3 = |e1 x e2| . face 3 and face 3p (B.5.b.4)

In (A.3.3) it is shown that E1 = det(R) e2 x e3 so that e2 x e3 lines up with E1 if det(R) > 0. Regardless of
^ directions. Thus,
the sign of det(R), we define the vector areas to point in the +E n

^
A1 ≡ |e2 x e3| E face 1p out-facing
1

A ≡ |e x e | E ^ face 2p out-facing
2 3 1 2
^ .
A3 ≡ |e1 x e2| E face 3p out-facing (B.5.b.5)
3

These three equations can be combined into the following single formula

^ .
An = |e1 x ... x e3| E // en missing (B.5.b.6)
n

where the ei are reordered for free due to the absolute value signs. But (A.6.2) says

En = det(R) (-1)n-1 e1 x ... x e3 // en missing (A.6.2) (B.5.b.7)

so
|En| = | det(R) | | e1 x ... x e3 | . // en missing (B.5.b.8)

Thus,

232
 Appendix B: Geometry of N-pipeds

^ |E | / |det(R)| = |det(S)| E
An = E n n n
n-1
= |det(S)| det(R) (-1) e1 x ... x e3 // en missing
n-1
= σ (-1) e1 x ... x e3 // en missing
where σ ≡ sign[det(S)] = sign[det(R)] (B.5.b.9)

To summarize, for N=3 one has

An = |det(S)| En = σ (-1)n-1e1 x ... x e3 // en missing

= σ (-1)n-1 Πxi≠n ei (B.5.b.10)

where the last line uses the shorthand notation of (A.10.5). This An result has the same formal form as
that of the 2-piped case (B.5.a.8).

Volume. The volume of a 3-piped (with each of the above An in turn treated as the base) is base times
height, so

volume(3) = | A1 • e1 | = | A2 • e2 | = | A3 • e3 |
or
volume(3) = | e2 x e3 • e1 | = | e3 x e1 • e2 | = | e1 x e2 • e3 | . (B.5.b.11)

Here is a drawing showing the last case (σ = +1), where "base" is A3 = | e1 x e2 | and "height" is e3 cosθ ,

(B.5.b.12)

Using ε notation one can write

e3 • e1 x e2 = (e3)i εijk(e1)j(e2)k = εijk (e1)j(e2)k (e3)i = εjki (e1)j(e2)k(e3)i

= det [ e1, e2, e3] = det(S) // see (3.2.7) (B.5.b.13)

so that

volume(3) = | e3 • e1 x e2 | = | εabc (e1)a(e2)b(e3)c | = | det [ e1, e2, e3] | = | det(S) | . (B.5.b.14)

This volume expression has the same form as that for the 2-piped, (B.5.a.10).

233
 Appendix B: Geometry of N-pipeds

(c) The 4-piped

Face Area: The faces of a 4-piped are 3-pipeds. For N=3, the 3-piped volume is stated in (B.5.b.14),

volume(3) = | εabc (e1)a(e2)b(e3)c | (B.5.c.1)

where e1,e2,e3 are 3D vectors. For the 3-piped which is "face 4" of the 4-piped -- a "near" face which
touches the origin of the 4D skewed en coordinate system -- vectors e1,e2,e3 are 4D vectors. The first 3
components of each of these 4D vectors are the same as the components of the 3D ei vectors, while the
4th components are all 0. This is so because face 4 lies in a plane defined by this 4th component being 0.
The volume(3) formula expressed in terms of these new 4D vectors is therefore | εabc4 (e1)a(e2)b(e3)c |,
where ε is now a 4D ε tensor. The conclusion is that

A4 = | εabc4 (e1)a(e2)b(e3)c | (B.5.c.2)

and this then is the scalar area of both face 4 and its partner face 4p, the far face. Similar arguments would
then support these other area expressions

A1 = | ε1abc (e2)a(e3)b(e4)c |
A2 = | εa2bc (e3)a(e4)b(e1)c |
A3 = | εab3c (e4)a(e1)b(e2)c | . (B.5.c.3)

Since indices on ε can be swapped for free due to the absolute value, the non-summed index can be put
first in all three cases and one may then conclude that

A1 = |e2 x e3 x e4| face 1 and face 1p

A2 = |e3 x e4 x e1| face 2 and face 2p
A3 = |e4 x e1 x e2| face 3 and face 3p
A4 = |e1 x e2 x e3| face 4 and face 4p . (B.5.c.4)

where, as discussed in (A.4.1) and (A.4.2),

Q=AxBxC is defined by Qk = εkabc AaBbCc . (B.5.c.5)

In (A.6.2) it was shown that E1 = det(R) e2 x e3 x e4 so that e2 x e3 x e4 lines up with E1. Regardless of
^ directions. Thus
the sign of det(R), we define the vector areas to point in the +E n

^
An = |e1 x ... x e3| E // en missing n = 1,2,3,4 (B.5.c.6)
n

where the ei are reordered for free due to the absolute value signs. But (A.6.2) says

En = det(R) (-1)n-1 e1 x ... x e4 // en missing (B.5.c.7)

so
|En| = | det(R) | | e1 x ... x e4 | . // en missing (B.5.c.8)

234
 Appendix B: Geometry of N-pipeds

Thus,
^ / |det(R)| = |det(S)| E
An = |En| En n
n-1
= |det(S)| det(R) (-1) e1 x ... x e4 // en missing
= σ (-1)n-1e1 x ... x e4 // en missing σ ≡ sign[det(S)] = sign[det(R)] (B.5.c.9)

To summarize, for N=4 one has

An = |det(S)| En = σ (-1)n-1e1 x ... x e4 // en missing

= σ (-1)n-1 Πxi≠n ei σ ≡ sign[det(S)] = sign[det(R)] (B.5.c.10)

This An expression has the same form as those of the 2-piped and the 3-piped found earlier.

Volume. The volume of a 4-piped (with each of the above An in turn treated as the base) is base times
height, so

volume(4) = | A1 • e1 | = | A2 • e2 | = | A3 • e3 | = | A4 • e4 | (B.5.c.11)
or
volume(4) = | e2 x e3 x e4 • e1 | = | e3 x e4 x e1 • e2 | = | e4 x e4 x e1 • e3 | = | e1 x e4 x e2 • e4 | .

Using ε notation one can write the first case as

e2 x e3 x e4 • e1 = (e1)a εabcd(e2)b(e3)c(e4)d = εabcd(e1)a(e2)b(e3)c(e4)d

= det [ e1, e2, e3, e4] = det(S) // see (3.2.7) (B.5.c.12)

so that

volume(4) = | εabcd(e1)a(e2)b(e3)c(e4)d | = | det [ e1, e2, e3, e4] | = | det(S) | (B.5.c.13)

This volume expression has the same form as those of the 2-piped and the 3-piped.

(d) The N-piped

Face Area: The faces of an N-piped are (N-1)-pipeds. If there had been an (N-1)-piped section prior to
this one, the volume formula there would have been

volume(N-1) = | εabc...x (e1)a(e2)b... (eN-1)x | (B.5.d.1)

where e1,e2,...eN-1 were (N-1)D vectors (that is, 3D vectors if N=4). For the N-piped which is "face N" of
the N-piped -- a "near" face which touches the origin of the ND skewed en coordinate system -- vectors
e1,e2,...eN-1 are ND vectors. The first N-1 components of each of these ND vectors are the same as the
components of the (N-1)D ei vectors, while the Nth components are all 0. This is so because face N lies
in a plane defined by this Nth component being 0. The volume(N-1) formula expressed in terms of these

235
 Appendix B: Geometry of N-pipeds

new ND vectors is therefore | εabc...xN (e1)a(e2)b... (eN-1)x |, where ε is now an ND ε tensor. The
conclusion is that ( compare this to the N=4 expression (B.5.c.2) ),

AN = | εabc...xN (e1)a(e2)b... (eN-1)x | (B.5.d.2)

and this then is the scalar area of both face N and its partner face Np, the far face. Similar arguments
would then support similar expressions for the other faces, for example,

A1 = | ε1abc...x (e2)a(e3)b... (eN-1)w (eN)x |

A2 = | εa2bc...x (e3)a(e4)b... (eN)w (e1)x | . (B.5.d.3)

Since indices on ε can be swapped for free due to the absolute value, the non-summed index can be put
first in all three cases and one may then conclude that

A1 = |e2 x e3 x e4 x e5... x eN| face 1 and face 1p

A2 = |e3 x e4 x e5 x e6... x e1| face 2 and face 2p
A3 = |e4 x e5 x e6 x e7... x e2| face 3 and face 3p
...
AN = |e5 x e6 x e7 x e8.. x e3| face N and face Np (B.5.d.4)

where, as discussed in (A.4.1) and (A.4.2),

Q = A x B x C.... X is defined by Qk = εkabc...x AaBbCc .....Xx . (B.5.d.5)

In (A.6.2) it was shown that E1 = det(R) e2 x e3 ... eN so that e2 x e3 ... eN lines up with E1. Regardless of
^ directions. Thus
the sign of det(R), we define the vector areas to point in the +E n

^
An = |e1 x ... x eN| E // en missing n = 1,2,3...N (B.5.d.6)
n

where the ei are reordered for free due to the absolute value signs. But (A.6.2) says

En = det(R) (-1)n-1 e1 x ... x eN // en missing (B.5.d.7)

so
|En| = | det(R) | | e1 x ... x eN | // en missing (B.5.d.8)

Thus,

^ / |det(R)| = |det(S)| E
An = |En| En n
n-1
= |det(S)| det(R) (-1) e1 x ... x eN // en missing
= σ (-1)n-1e1 x ... x eN // en missing σ ≡ sign[det(S)] = sign[det(R)] (B.5.d.9)

To summarize, for N=N one has

An = |det(S)| En = σ (-1)n-1e1 x ... x eN // en missing

= σ (-1)n-1 Πxi≠n ei σ ≡ sign[det(S)] = sign[det(R)] (B.5.d.10)

236
 Appendix B: Geometry of N-pipeds

This An expression has the same form as those of the 2-piped, the 3-piped and the 4-piped.

Volume. The volume of a N-piped (with each of the above An in turn treated as the base) is base times
height, so

volume(N) = | A1 • e1 | = | A2 • e2 | = ... = | AN • eN |
or
volume(N) = | e2 x e3 x e4....eN • e1 | = ... (B.5.d.11)

Using ε notation one can write the first case as

e2 x e3 x e4...eN • e1 = (e1)a εabc...x(e2)b(e3)c.....(eN)x = εabc...x(e1)a(e2)b(e3)c....(eN)x

= det [ e1, e2, e3, ....eN] = det(S) // see (3.2.7) (B.5.d.12)

where x is the Nth letter of the alphabet, so that

volume(N) = | εabc...x(e1)a(e2)b(e3)c....(eN)x | = | det [ e1, e2, e3, ....eN] | = | det(S) | . (B.5.d.13)

This volume expression has the same form as those of the 2-piped, the 3-piped and the 4-piped. This
result is also consistent with the volume(N-1) expression stated in (B.5.d.1).

B.6 Summary of Main Results of this Appendix

1. One way to write the equation of a plane in N dimensions is

r•^
p=p r = (x1, x2, .....xn) (B.6.1)

where ^p is the unit vector normal to the plane which points "away from the origin", and where p > 0 is the
distance of closest approach of the plane to the origin. In the limit p→0, the plane passes through the
origin and the equation is then r•^
p = 0 where ^p is either normal to the plane.

2. An N-piped has 2N vertices as shown by the inductive construction method presented above. (B.6.2)

3. The locus of points making up the (closed) interior of an N-piped spanned by e1...eN is given by

rvolumeN = Σn=1Nαnen 0 ≤ αn ≤ 1 (B.6.3)

The tails of all the vectors e1...eN meet at the origin of RN space.

4. There are N pairs of faces on an N-piped, and each face is an (N-1)-piped having 2N-1 vertices. The
total face count is 2N. Each face is spanned by a subset of N-1 of the base vectors en, so each face is

237
 Appendix B: Geometry of N-pipeds

"missing" one of the en and the face is labeled using the index of this missing base vector. The loci of
points making up the faces of an N-piped are given by

rface(i) = Σn≠iαnen 0 ≤ αn ≤ 1 i = 1,2...N

rface(ip) = Σn≠iαnen + ei 0 ≤ αn ≤ 1 i = 1,2...N (B.6.4)

where "face i" has a corner touching the origin of the N-piped (near face), while its parallel partner face
"ip" does not touch the origin (far face).

5. If the N-piped spanning vectors en are the tangent base vectors associated with some transformation F,
then Ei• ej = δi,j where Ei are the reciprocal base vectors. In this case, the equations of the planes of
the faces of the N-piped can be written in the form shown in item 1 above,

^ •r
E =0 ^ •r
E = 1/Ei i = 1,2...N (B.6.5)
i (face i) i (face ip)

^
so that both faces of a pair i are planar (in N dimensional space) and they have the same normal vector E i
so the faces of a pair lie on parallel planes.

6. The vector Ei is an outward-facing normal for face ip, while -Ei is an outward-facing normal vector
for face i (which touches the origin). (B.6.6)

7. The out-facing vector area of face ip of an N-piped can be expressed as

Ai = |det(S)| Ei

Ai = σ (-1)i-1 Πxj≠i ej σ ≡ sign[det(S)] = sign[det(R)]

Ai = σ (-1)i-1 e1 x e2 ... x eN // ei missing (B.6.7)

where ei is the vector missing from the face's spanning set. The outfacing area for face i is - Ai. The last
two lines are shorthands for the following, as discussed in (A.5.2),

(Ai )α = σ (-1)i-1 εαabc..x (e1)a (e2)b ... (eN)x // where (ei)i and i are missing (B.6.8)

For N=2 the last two expressions for Ai are interpreted as shown in (B.5.a.9),

Ai = σ (-1)i-1 e3-i x ^
3 i = 1,2 (B.6.9)

8. The volume of an N-piped spanned by e1...eN is given by

volume(N) = | det(S) | = | det [ e1, e2, e3, ....eN] | = | εabc...x(e1)a(e2)b(e3)c....(eN)x | (B.6.10)

where one can regard the tangent base vectors as the columns of the linearized transformation matrix S as
shown in (3.2.7).

238
 Appendix C: Elliptic Polar Coordinates

Appendix C: Elliptical Polar Coords, Views of x'-space, Jacobian Integration Rule

This Appendix is written in the developmental notation of Chapters 1-6.

C.1 Elliptical polar coordinates

The 2D "elliptic" coordinate system has coordinate lines which are orthogonal ellipses and hyperbolas.
When rotated about its two symmetry axes, this system generates 3D prolate or oblate spheroidal
coordinates. This is not the 2D coordinate system described in this Appendix. For "elliptical polar"
coordinates, the coordinate lines are taken instead as the ellipses from elliptic coordinates, and the rays
from polar coordinates. This non-orthogonal system is perhaps not very useful, but provides a good
"sandbox" in which to study general aspects of coordinate systems.
The transformation x' = F(x) is given by
x'-space x-space (Cartesian)
2 2 2 2 2 2 2 2
ρ = x /a + y /b x + y = r still x'1 = θ x 1= x
tanθ = y/x x'2 = ρ x2 = y (C.1.1)

Writing the first equation above as

1 = x2/(ρa)2 + y2/(ρb)2 (C.1.2)

it should be clear that ρ serves to label an ellipse of semi-major axis ρa, and semi-minor axis ρb, while θ
labels the ray at angle θ, as in polar coordinates. The inverse transform x = F-1(x') is given by

x = aρcosθ x/a = ρcosθ ⇒ x2/a2 + y2/b2 = ρ2

y = bρsinθ y/b = ρsinθ ⇒ tanθ = y/x . (C.1.3)

The matrix S is given by

S11 = (∂x/∂θ) = -aρsinθ

S12 = (∂x/∂ρ) = acosθ Sik ≡ ( ∂xi/∂x'k) (2.1.5)
S21 = (∂y/∂θ) = bρcosθ
S22 = (∂y/∂ρ) = bsinθ (C.1.4)

-aρsinθ acosθ -sinθ/(aρ) cosθ/(bρ) ⎞

S = ⎛⎝ bρcosθ bsinθ ⎞⎠ ⇒ det(S) = -abρ and R = S-1 = ⎛⎝ cosθ/a sinθ/b ⎠ .

The tangent base vectors en can be read off as the columns of S, see (3.2.7) :

e1 = ρ(-asinθ,bcosθ) = eθ |eθ| = ρ a2sin2θ + b2cos2θ ≡ hθ eθ = |eθ| ^

eθ
e2 = (acosθ,bsinθ) = eρ |eρ| = a2cos2θ + b2sin2θ ≡ hρ eρ = |eρ| ^
eρ . (C.1.5)

The covariant metric tensor from (5.7.9) or (5.11.3) is,

239
 Appendix C: Elliptic Polar Coordinates

2 2 2 2 2 2 2
ḡ' = STS =
⎛ρ {a 2sin2 (θ) + b cos (θ)} [b2 -a 2]ρ sin(θ)cos(θ)
2 2
⎞ = ⎛⎜ e1•e1 e1•e2 ⎞⎟ (C.1.6)
⎝ [b -a ]ρ sin(θ)cos(θ) a cos (θ) + b sin (θ)⎠ ⎝ e2•e1 e2•e2 ⎠

which is clearly non-diagonal (but symmetric) as expected. When a = b = 1 it reduces to the polar
coordinates system metric tensor where then ρ = r. The coordinate system is non-orthogonal because
e1•e2 ≠ 0, or equivalently, because ḡ' is non-diagonal.

C.2 Forward coordinate lines

Here is a Maple plot of some x-space forward coordinate lines (parameters a = 2 and b = 1)

(C.2.1)
where ρ is the x'-space vertical axis. The coordinate lines in x-space are plotted using these equations,

y = b ρi2-(x/a)2 // ellipses ρi= 1,2...10 10 ellipses

y = x tanθi // rays θi = 2π (i/20) , i = 1,2...20 20 rays (C.2.2)

which are obtained from the forward transformation equations

ρ2 = x2/a2 + y2/b2
tanθ = y/x . (C.1.1)

C.3 Inverse coordinate lines

Here is a Maple plot of some x'-space inverse coordinate lines (parameters a = 2 and b = 1)

240
 Appendix C: Elliptic Polar Coordinates

The coordinate lines in x'-space are plotted using these equations (C.3.1)

ρ = xi/(acosθ) xi = -10 to +10 20 blue curves

ρ = yi/(bsinθ) yi = -10 to +10 20 red curves (C.3.2)

which are obtained from the inverse transformation equations

x = aρ cosθ
y = bρ sinθ (C.1.3)

The secθ and cscθ curve families appear to "change shape", but that is just what happens when functions
are scaled up vertically but not horizontally. If one plots one sine hump at different vertical scalings, the
humps have different shapes.

C.4 Drawing a contravariant vector V in x-space: the meaning of V'n .

A contravariant vector field V(x) can be expanded in these two ways, as shown in (6.6.9),

V = V1(x) ^
1 + V2(x) ^
2 = Vx(x) ^
x + Vy(x) ^
y // un = ^
n for Cartesian

V = V'1(x') e1 + V'2(x') e2 = V'θ(x') eθ + V'ρ(x') eρ V'(x') = R(x) V(x) (C.4.1)

where the V'n are the components of V transformed into x'-space where V becomes V'. The prime is not
necessary on V'ρ but we maintain it as a reminder that it is an x'-space component. R(x) is the matrix of
(2.1.6) and en are the tangent base vectors of Chapter 3. The fields V'n(x') are "the components of vector
field V' in x'-space", since V' = RV , or V'i = RijVj. Moreover, these V'n(x') are "expressed in terms of
curvilinear coordinates" x'. If one is asked to "express a vector V in curvilinear coordinates", one is
usually being asked to write V as the second expansion above. The vectors en and V exist in x-space, and
in the second expansion it just happens that the coefficients V'n(x') are the components of V', the
transformed vector in x'-space, when it is expanded on the axis-aligned vectors e'n in x'-space.

241
 Appendix C: Elliptic Polar Coordinates

Here is a graphical representation of this vector V in x-space:

(C.4.2)

As advertised, the tangent base vectors are not at right angles. The V parallelogram accurately illustrates
the equation V = V'θ eθ + V'ρ eρ. Since x-space is Cartesian, there is no distinction between Cartesian
length (graphical length) and covariant length for vectors in x-space. Graphically, one could find the
values for V'θ and V'ρ as follows: (1) for the point (x,y), compute the vectors eθ and eρ and compute their
lengths |eθ| = h'θ and |eρ| = h'ρ ; (2) draw the parallelogram shown aligned with these vectors for some
given V and find the edge lengths. The edges of the parallelogram are V'θ h'θ and V'ρ h'ρ so then the
values of V'θ and V'ρ can be found.
The alternative method is to compute R and use V'i = RijVj.

C.5 Drawing a contravariant vector V' in x'-space: two "Views"

As with previous examples, the drawing (C.4.2) is drawn to the right of an x'-space drawing as follows:

(C.5.1)
In Chapter 3 the axis-aligned basis vectors e'n were introduced as

e'n , n = 1,2...N // (e'n)i = δn,i e'1 = (1,0,0...) etc (3.2.1)

and en was shown to be a contravariant vector,

242
 Appendix C: Elliptic Polar Coordinates

e'n = R(x) en . (3.3.2)

Applying matrix R(x) to the equation V = V'θ eθ + V'ρ eρ one gets the expansion shown in Fig (C.5.1),

V' = V'θ e'θ + V'ρ e'ρ (6.6.15) (C.5.2)

which appears first in the list of expansions of V' in (6.6.15). There is no ambiguity concerning this last
equation. Ambiguity can arise, however, when one tries to represent this equation graphically in x'-space.
There are two very different "views" one can take of a drawing in x'-space. In the first view, we take
x'-space to be "flat" (Cartesian) so that g' = 1. In the second view, we take x'-space to be "curved" with g'
≠ 1. These views are really two different x'-spaces since the metric tensors are different.

Cartesian View of x'-space

(C.5.3)

In this view of x'-space the length (norm) of a vector is given by |A|2 = δijAiAj = ΣAi2, so one has,
since (e'n)i = δn,i,

ḡ' = 1 |e'n| = 1 e'n = ^

e 'n = ^
n' = the usual axis-aligned unit vectors in x'-space

V' = V'θ ^
θ + V'ρ ^
ρ ^
θ=^
e '1 ^
ρ=^
e '2 . (C.5.4)

The left-side of (C.5.1) is, in this view, a "normal Cartesian graph" and the vectors add up properly. For
example Pythagoras tells us that

|V'|2 = V'θ2 + V'ρ2

and
^
θ•^
θ =1
^ ^
ρ•ρ =1
^
θ•^
ρ=0 . (C.5.5)

243
 Appendix C: Elliptic Polar Coordinates

This Cartesian View, which is x'-space with ḡ' = 1, is appropriate in applications in which it is not
required or desired that norms and dot products be tensorial scalars, as discussed at the end of Section 5.2.
For example, when ḡ' is set to 1, one has |V'| ≠ |V| in (C.5.1).
Another use of this view involves integration as will be seen below.

Curvilinear View of x'-space

(C.5.6)

In this view of x'-space, one assumes that ḡ' takes a value which enforces the scalarity of norms and dot
products between x-space and x'-space, which is to say, one takes ḡ' = ST ḡ S from (5.7.6) where ḡ is the
x-space metric tensor for x-space. Normally ḡ=1 (Cartesian x-space), so ḡ' = STS. In this Curvilinear
View, then, the length |A'| of a contravariant vector A' is determined by |A'|2 = ḡ'ijA'iA'j where

ḡ' = STS ≠ 1 |e'n| = |en| = h'n ≡ ḡ 'nn n = 1,2 for θ,ρ

(C.5.7)

V' = V'θ e'θ + V'ρ e'ρ = (V'θ h'θ) ^

e 'θ + (V'ρ h'ρ) ^
e 'ρ ^
e 'n ≡ e'n/ |en| = e'n/ h'n

Vx2 + Vy2 = |V|2 = |V'|2 ≠ (V'θ h'θ)2 + (V'ρ h'ρ)2 // unless x'i are orthogonal coordinates

This last inequality says that in the Curvilinear View the Pythagorean Theorem is invalid. In fact

|V'|2 = ḡ'ijV'iV'j = ḡ'θθ V'θ2 + ḡ'ρρ V'ρ2 + 2 ḡ'θρV'θ V'ρ

= (V'θ h'θ)2 + (V'ρ h'ρ)2 + 2 ḡ'θρV'θ V'ρ . (C.5.8)

In writing |e'n| = |en| and |V'|2 = |V|2 above, we use the rule shown in (5.10.4) which says |A'|2 = |A|2 for
any contravariant vector A (|A|2 is a scalar ). Moreover,

244
 Appendix C: Elliptic Polar Coordinates

^
e 'θ • ^
e 'θ = e'θ • e'θ / (h'θ2) = eθ • eθ / (h'θ2) = ḡ'θθ / (h'θ2) = 1
^
e 'ρ • ^
e 'ρ = e'ρ • e'ρ / (h'ρ2) = eρ • eρ / (h'ρ2) = ḡ'ρρ / (h'ρ2) = 1
^
e 'θ • ^
e 'ρ = e'θ • e'ρ / (h'θh'ρ) = eθ • eρ / (h'θh'ρ) = ḡ'θρ / (h'θh'ρ) ≠ 0 <= !! (C.5.9)

so the ^e 'n are unit vectors having unit covariant length, but ^ e 'θ • ^
e 'ρ ≠ 0 despite the fact that these
vectors are drawn at right angles in the x'-space graph above, en = h'n ^ e 'n. One might imagine trying to
slant the lines in (C.5.6) to cause all intersection points to have angles which match the metric tensor,
which is to say, at each intersection point one would need an angle ψ where ^ e 'θ • ^
e 'ρ = cosψ. But in
^ ^
general e 'n • e 'm = ḡ'nm / (h'nh'm) has a different value at every point, so such a graph would be quite
complex.
The upshot is that for a non-orthogonal system, the axes in x'-space are still drawn at right angles and
the purpose of the graph is mainly to "locate" all the points x' which correspond to points x in x-space
according to x' = F(x). The graph does successfully represent the idea that V' = V'ρ e'ρ + V'θ e'θ, but one
must give up on Euclidean geometry for this vector sum triangle. It might be imagined that the x'-space
graph is the projection onto the plane of paper of some vectors drawn on a curved surface emerging from
the plane of paper, and that is then why Pythagoras is wrong.

In the case of an orthogonal coordinate system (diagonal ḡ'), the 90 degree angles between the axes in x'-
space are accurate representations of the fact that ^e 'n• ^
e 'm = 0 when n ≠ m. And since scalars are
preserved, one has in the Curvilinear View,

|V'|2 = Σn (h'nV'n)2 = Σn V'n2 = |V|2 = Σn Vn2 // orthogonal only

where
V'n ≡ h'nV'n and V' = Σn V'n ^
e 'n
and V = Σn V'n ^
en (C.5.10)

One can then still apply regular Euclidean geometry to the vector addition N-piped in x'-space in the
sense that |V'|2 = Σn (h'nV'n)2.
Although the x'-space perpendicular-unit-vectors and Pythagorean paradoxes go away if the x'
coordinates are orthogonal (as they would be for polar coordinates where a = b = 1), one still has the two
x'-space views to keep in mind: the Cartesian View with ḡ' = 1 and the Curvilinear View with ḡ' = STS.

C.6 Drawing the specific contravariant vector dx in x-space and x'-space

Since dx is the primordial contravariant vector, everything stated in the last two Sections applies with V
→ dx and V'θ → dx'θ = dθ, V'ρ → dx'ρ = dρ, where we finally drop the primes on dθ and dρ. The
expansions of dx and dx' are,

dx = dθ eθ + dρ eρ // in x-space
dx' = dθ e'θ + dρ e'ρ // in x'-space (C.6.1)

For V = dx Fig (C.5.1) becomes,

245
 Appendix C: Elliptic Polar Coordinates

(C.6.2)

It must be understood that now the vector arrows like dx are highly magnified and in reality are very
small compared to, say, the curvature of the ellipse. From above,

e'θ • e'θ = ḡ'θθ ^

e 'θ • ^
e 'θ = 1
e'ρ • e'ρ = ḡ'ρρ ^
e 'ρ • ^
e 'ρ = 1
e'ρ • e'θ = ḡ'ρθ ^
e 'θ • ^
e 'ρ = ḡ'θρ / (h'θh'ρ) (C.6.3)

and once again the "right angle" in the x'-space picture is deceptive.
The x'-space side of (C.6.2) is subject to the two "views" described above:

Cartesian View of x'-space: (elliptical polar coordinates)

ḡ' = 1 |e'n| = 1
dx' = dθ ^
θ + dρ ^
ρ ^
θ=^
e '1 ^
ρ=^
e '2
^ ^
θ•θ =1 ^
ρ•^ ρ =1 ^
θ•^
ρ=0 . (C.6.4)

Curvilinear View of x'-space: (elliptical polar coordinates)

ḡ' = STS |e'n| = |en| = h'n ≡ ḡ 'nn

dx' = dθ e'θ + dρ e'ρ = (dθ h'θ) ^

e 'θ + (dρ h'ρ) ^
e 'ρ

(dx)2 + (dy)2 = |dx|2 = |dx'|2

= (dθ h'θ)2 + (dρh'ρ)2 + 2 ḡ'θρdθdρ

= (dθ h'θ)2 + (dρ h'ρ)2 only if ḡ'θρ = 0 (orthogonal) . (C.6.5)

246
 Appendix C: Elliptic Polar Coordinates

Just to have a specific orthogonal coordinates example, we reduce (C.1.3) to polar coordinates by setting
ρ2 0 ⎞
a = b = 1. In this case, (C.1.6) reduces to ḡ' = ⎛
⎝ 0 1 ⎠ as in (5.13.11) so hθ = ρ, hρ = 1 and ḡ'θρ = 0.
We can then restate the two Views in this case

Cartesian View of x'-space: (polar coordinates)

ḡ' = 1 |e'n| = 1
dx' = dθ ^
θ + dρ ^
ρ ^
θ=^
e '1 ^
ρ=^
e '2
^
θ•^ θ =1 ^
ρ•^ ρ =1 ^
θ•^
ρ=0 .

(ds)2 = |dx|2 = dx • dx = (dx)2 + (dy)2 // from x-space (which has ḡ = 1 )

(ds')2 = |dx'|2 = dx' • dx' = (dθ)2 + (dρ)2 ≠ |dx|2 = (ds)2 (C.6.6)

Curvilinear View of x'-space: (polar coordinates)

ρ2 0 ⎞
ḡ' = STS = ⎛ |e'θ| = |eθ| = h'θ = ρ ≡ ḡ 'θθ
⎝ 0 1⎠
|e'ρ| = |eρ| = h'ρ = 1 ≡ ḡ 'ρρ

dx' = dθ e'θ + dρ e'ρ = (dθ ρ) ^

e 'θ + (dρ) ^
e 'ρ

(ds)2 = |dx|2 = dx • dx = (dx)2 + (dy)2 // from x-space (which has ḡ = 1 )

(ds')2 = |dx'|2 = dx' • dx' = (dθ ρ)2 + (dρ)2 = |dx|2 = (ds)2 (C.6.7)

C.7 Study of how dx transforms in the mapping between x-space and x'-space

Consider this drawing which shows a representative set of vectors dx in x-space (the bars), along with the
forward mappings (dx' = F(dx) or dx' = Rdx ) of the corresponding vectors dx' in x'-space. The vectors on
the right all point up, those on the left point generally to the northeast.

247
 Appendix C: Elliptic Polar Coordinates

(C.7.1)

Now select the red dx bar on the right and take it to be the dx of Fig (C.6.2). First determine the tangent
base vectors eθ and eρ at the location of the red bar. Then setting dx = dθ eθ + dρ eρ, consider the value of
the two numbers dθ and dρ for this red bar. Graphically, knowing which way eθ and eρ point at the
bottom of the red dx, one expects dθ > 0 and dρ > 0. The red dx' bar on the left has these Cartesian values
dθ and dρ, and has a Cartesian-view length of |dx|2 = (dθ)2+(dρ)2. One can see from the picture that these
Cartesian lengths vary for the 10 bars shown, though the lengths are all the same in x-space. The
Curvilinear-view lengths of the x'-space bars are all the same, and are equal to the Cartesian length of
those bars in x-space since dx'•dx' = dx•dx.
Consider now some bar mapping in the other direction:

(C.7.2)

Now the dx bars on the right all have different lengths. Those on the left have the same Cartesian length,
which is what the drawing shows, but each one's Curvilinear-view length matches that of its
corresponding bar on the right. The ratio of the length of a bar on the right to the Cartesian length of the
corresponding bar on the left is the scale factor hθ which recall is a function of location in space:

248
 Appendix C: Elliptic Polar Coordinates

bar on right = dx(1) = e1 dx'1 = ^e 1 h'1 dx'1 = eθ dθ = ^ e θ hθ dθ graph length = hθ dθ

bar on left (Cartesian view) = dx'(1)
= e'1 dx' = ^
1
e '1 dx' = ^
1
e θ dθ graph length = dθ

⇒ right bar length / left bar length = hθ = ρ a2sin2θ + b2cos2θ (increases with ρ) (C.7.3)

If a=b, then hθ = ρ and the bar length on the right is then ρdθ as is obvious in polar coordinates.

C.8 A Derivation of the Jacobian Integration Rule

Consider now an integral ∫dθdρ f(θ,ρ). The tiny rectangles of area dθdρ, like the specific gray and orange
ones highlighted on the left above, are regarded for the purposes of integration as being in the Cartesian
view of x'-space. One then writes [ dA' is called dV' in Chapter 8 ]

dA' ≡ dρdθ = the area of a differential patch in Cartesian-view x'-space (C.8.1)

This is the graphical area one sees in the picture. There is no need to define or consider any Curvilinear-
view area in x'-space because the Cartesian-view area is being used.
In the limiting process which defines the integration, each dθdρ patch on the left has the same area
dρdθ. The interior of each patch on the left maps into some parallelogram patch on the right. One is not
surprised to see that the patch areas on the right are different, though they map into patches on the left of
the same Cartesian-view area. As shown in (8.4.d.3) the ratio of the two patch areas (volumes) is the
absolute value of the Jacobian |J(x')|,

(area of skewed patch on the right at location x) = |J(x')| dA' = |J(x')| dρdθ (C.8.2)

This is not what we mean by "the Jacobian Integration Rule" in the Section title. That is coming below
and it is going to involve the quantity dxdy.
The mapping shown above between patches is an N=2 example of the general N-dimensional
discussion in Section 8.2 which describes an orthogonal differential N-piped in (Cartesian-view) x'-space
mapping into a non-orthogonal differential N-piped in x-space.
Now back to the integration issue. There are two ways an integration can be done in Cartesian x-
space:

integral of f(x) = lim Σi dA1(xi) f(xi) dA1(xi) = patches shown on the right above

integral of f(x) = lim Σi dA2(xi) f(xi) dA2(xi) = dxdy (C.8.3)

In the first integral, every patch dA1(xi) on the right has a different shape and a different area as the
integral is computed in the usual limiting-sum manner. The gray and orange patches on the right are two
of these many patches. Despite their non-uniform shape and area, this rag-tag band of patches certainly
"covers" the area being integrated over, and does so perfectly in the calculus limit. The area of one of
these rag-tag patches is |J(x')|dA' = |J(x')|dθdρ and the areas are different because the Jacobian is a
function of x = x(x').

249
 Appendix C: Elliptic Polar Coordinates

In the second integral, every patch dA2(xi) has the same area dxdy, so really dA2(xi) does not depend on
xi in this form of the integration. One such dxdy patch is shown in green above. The coverage of the dA2
patches is of course also "perfect coverage" in the calculus limit.

Since both integrals cover the same area perfectly, they both give the same result in the limiting process
that defines the integral. This point is sometimes misunderstood. One is not just "replacing" a
parallelogram patch such as the orange one on the right with some dxdy patch that approximates it in
area, like the green patch. The statement is about an integration. Thus one has

lim Σi dA1(xi) f(xi) = lim Σi dA2(xi) f(xi) (C.8.4)

or
∫[|J(x')| dθdρ] f(x(x')) = ∫[dxdy] f(x) (C.8.5)

where on the left f(x) = f(x(x')) where x = F-1(x') ≡ x(x'). In the sense of distribution theory (Stakgold
Chapters 1 and 5), one can then make this symbolic statement

|J(θ,ρ)| dρdθ = dxdy (C.8.6)

where the meaning of this symbolic equality is the integral statement above,

∫D dxdy f(x) = ∫D' dθdρ |J(x')| f(x(x')) , (C.8.7)

valid for any integrable f(x) and any integration region D (region D' corresponds to D in x'-space.) Either
of these last two equations constitute the "Jacobian Integration Rule" of the Section title.

The integral on the left is well defined in 2D calculus, so the expression on the right shows how to
"evaluate the integral on the left in curvilinear coordinates".

At this point one may introduce a new but obvious symbol

dA ≡ dxdy (C.8.8)

so the above equality of integrals can be written

∫dA f(x) = ∫ dA' |J(x')| f(x(x')) |J(x')| dA' = dA (C.8.9)

In N dimensions, dA and dA' are differential "volumes", and the general Jacobian Integration rule takes
the form,

250
 Appendix C: Elliptic Polar Coordinates

∫dV f(x) = ∫ dV' |J(x')| f(x(x')) |J(x')| dV' = dV

dV' ≡ dx'1dx'2....dx'N = the volume of an orthogonal differential N-piped

in the Cartesian-view x'-space

dV = dx1dx2....dxN = the volume of an orthogonal differential N-piped in x-space. (C.8.10)

Notice that these are not the two N-pipeds which "map into each other" as noted above. The N-piped dV
has nothing to do that that mapping which involved a non-orthogonal N-piped in x-space.

To finish off our sample N=2 case, recall from earlier that for our polar elliptical coordinate system

|J'(x')| = | det(S)| = abρ (C.8.11)

and therefore

∫dxdy f(x,y) = ∫ dθdρ |J(x')| f(aρ cosθ, bρ sinθ) = ab ∫dθdρ ρ f(aρ cosθ, bρ sinθ) . (C.8.12)

In the limit of regular polar coordinates, one then has a = b = 1 and ρ = r so

∫dxdy f(x,y) = ∫rdrdθ f(rcosθ, rsinθ) (C.8.13)

which is the familiar result.

251
 Appendix D: Tensor Densities

Appendix D: Tensor Densities and the ε tensor

In this Section, DN means Developmental Notation.

D.1 Definition of a tensor density

Picture A is used in this Appendix along with the Standard Notation of Chapter 7.

(D.1.1)

First, equations (5.12.20) summarize facts about the Jacobian J which were converted to Standard
Notation in (7.5.18) through (7.5.22). Here we add a symbol σ ≡ sign(J) and restate those facts in
Standard Notation as

J ≡ det(Sij) = σ sg' / sg = σ(sg'/sg)1/2 = σ(g'/g)1/2 ⇒ (g'/g)1/2 = σJ = |J| > 0 J2 = (g'/g)

s = sign[det(gij)] = sign(g) = sign(g') g = det(gij) sg = |g| > 0 Sij ≡ (∂xi/∂x'j)

σ = sign[det(Sij)] = sign(J) g' = det(g'ij) sg' = |g'| > 0 . (D.1.2)

For proper Lorentz transformations of special relativity, det(S) = 1 so σ = +1. For curvilinear coordinates,
one normally selects an ordering of the xi so that σ = +1, such as r,θ,φ in spherical coordinates.
Nevertheless, we allow for the possibility of J < 0.

Next, recall our generic sample tensor transformation (7.10.1), written two ways using (7.5.13) Sab=Rba :

T ' abcde = Raa' Rbb' Rcc' Sd'd Se'e Ta'b'c'd'e'

T ' abcde = Raa' Rbb' Rcc' Rdd' Ree' Ta'b'c'd'e' . (7.10.1) (D.1.3)

T is a mixed rank-5 tensor, meaning it transforms as shown above with respect to the underlying
transformation F. T is a regular standard-issue tensorial tensor.

Now suppose instead that the object T were to transform like this, with J being the Jacobian noted above,

T ' abcde = J-W Raa' Rbb' Rcc' Rdd' Ree' Ta'b'c'd'e' (D.1.4)

where the extra factor J-W has been introduced. If T transforms this way, it is called a tensor density of
weight W. Thus, an ordinary tensor is a tensor density of weight 0.
The convention for the sign of W used here is that of Weinberg p 99 Eq. (4.4.4), which equation has
the following factor on the right side of a sample tensor density transform equation,

252
 Appendix D: Tensor Densities

|∂x'/∂x|W ≡ [det(∂x'/∂x)]+W = [ det(∂x'i/∂xk) ]+W = J-W . (D.1.5)

Some authors use -W as the "weight" instead of +W, but we shall stick with Weinberg's convention.
An immediate example of a tensor density is provided by (g'/g)1/2 = |J| rewritten as

g' = J2 g = J-(-2) g ⇒ weight(g) = -2 g is a scalar density of weight - 2 . (D.1.6)

so g ≡ det(gij) is a scalar density of weight -2. Notice that from g one can construct other scalar densities
of other weights, for example :

g'-1 = J-(2) g-1 ⇒ weight(g-1) = +2 g'-1 is a scalar density of weight +2 . (D.1.7)

D.2 A few facts about tensor densities

1. It is pretty obvious that a sum of two index-similar tensor densities of weight W has weight W.
(D.2.1)

2. Contracting indices within a tensor density does not alter its weight W. (D.2.2)

If indices a and d are contracted in the (D.1.4) example above, one gets

T ' abcae = J-W Raa' Rbb' Rcc' Rad' Ree' Ta'b'c'd'e'

= J-W (Raa'Rad') Rbb' Rcc' Ree' Ta'b'c'd'e'

= J-W δa'd' Rbb' Rcc' Ree' Ta'b'c'd'e'

= J-W Rbb' Rcc' Ree' Ta'b'c'a'e' .

The factor J-W just sits there, impervious to contraction activities.

3. Going the other direction, when a larger tensor density is formed from two smaller ones, called a direct
product or outer product, the weights get added. (D.2.3)

Example 1: Start with two tensor densities A and B of weights W1 and W2, form the outer product:

A'a = J-W1 Raa'Aa'

B'cd = J-W2 Rcc'Rdd' Bc'd'

⇒ (A'a B'cd) = J-(W1+W2) Raa' Rcc'Rdd' (Aa' Bc'd') .

Example 2: Start with |g'| = J2|g| from (D.1.2) and raise both sides to the power -W1/2 :

|g'|-W1/2 = J-W1 |g|-W1/2 . // no R factors since scalar (D.2.3a)

253
 Appendix D: Tensor Densities

Thus, the factor ( |g|-W1/2) is a scalar density of weight W1. Let this be the first factor of an "outer
product" where the second factor is the following tensor density of weight W2,

B'cd = J-W2 Rcc'Rdd' Bc'd' // same as in Example 1

⇒ (|g'|-W1/2B'cd ) = J-(W1+W2) Raa' Rcc'Rdd' (|g|-W1/2 Bc'd') . (D.2.3b)

If one selects W1 = –W2, the added factor neutralizes the weight of the tensor density to which it is
prefixed, generating thereby a regular tensor (weight 0). So if tensor density B has weight W,

(|g'|W/2 B'cd) = Raa' Rcc'Rdd' (|g|W/2 Bc'd') (D.2.3c)

and then (|g|W/2 Bij) transforms under F as a regular tensor. (One should always keep in mind the fact
that there is an underlying transformation x' = F(x) upon which the House of Tensor is built ). This is a
standard method of converting a tensor density to a regular tensor.

4. Although sometimes authors take a differing stance for certain tensors like ε below, we shall assume
that indices are raised and lowered on a tensor density in exactly the same way they are raised and
lowered on an ordinary tensor of the same index structure (see Section 7.4). This means the gab raises an
index and gab lowers an index. (D.2.4)

5. Raising or lowering an index does not change the weight of a tensor density. (D.2.5)

Again, using our generic example (D.1.4) above,

T 'abcde = J-W Raa' Rbb' Rcc' Rdd' Ree' Ta'b'c'd'e' (1)

T 'abcde = g'ex T 'abcdx (2) // raise last index e on T' in x'-space

Ta'b'c'd'e' = ge'e" Ta'b'c'd'e" . (3) // lower last index on T in x-space

Therefore,

T 'abcde = g'ex [ J-W Raa' Rbb' Rcc' Rdd' Rxe' Ta'b'c'd'e'] // this is (2) + (1) above

= g'ex [ J-W Raa' Rbb' Rcc' Rdd' Rxe' (ge'e" Ta'b'c'd'e")] // use (3) above

= J-W Raa' Rbb' Rcc' Rdd' (g'ex Rxe' ge'e") Ta'b'c'd'e" // regroup

= J-W Raa' Rbb' Rcc' Rdd' (Ree") Ta'b'c'd'e" // from last line in (7.5.9)†

and again J-W passively watches all the action fly by. The weight of our generic tensor density with its last
index raised is still W. † The last step uses (7.5.9) Rab = g'aa'Ra'b'gb'b but we raise a, lower b, and
reverse two tilts to get Rab = g'aa'Ra'b'gb'b .

254
 Appendix D: Tensor Densities

6. The covariant dot product of vector densities A and B of weights W and w is a scalar density of
weight W + w and therefore A'•B' = J-(W+w) A•B . (D.2.6)

Proof: First form the rank-2 tensor density AiBj which by (D.2.3) has weight W+w. Lower the second
index and the mixed rank-2 tensor AiBj by item (D.2.5) still has weight W+w. Then contract to get A•B
= AiBi and by (D.2.2), the weight is still W+w.

Corollary: The magnitude of a vector density A of weight W is a scalar density of weight W, and
therefore |A|' = J-W |A| . (D.2.7)

Proof: |A|2 = A • A has weight 2W by (D.2.6), meaning |A'|2 = J-2W |A|2. Therefore |A'| = J-W |A| .

7. As J→1, tensor densities become true tensors. (D.2.8)

One could imagine some limiting/morphing process on an underlying transformation F such that the
linearized transformation matrix R approaches a rotation matrix at all points in space ( RRT= 1 and detR =
1, DN) and then J = detS → 1. In this case J-W → 1-W = 1 and therefore any tensor density, regardless of
its weight W, becomes an ordinary tensor. Perhaps we should restrict this comment to underlying
transformations F having J = detS > 0 since passing through detS = 0 is problematical. As was noted
below (5.12.16), for an invertible x' = F(x), one does not have detS(x) = 0. at any x.

Example: The cross product considered in Section D.7 below of N-1 contravariant vectors becomes in this
limit an ordinary covariant vector. If g=1 in x-space, then g' = RRT = 1 in x'-space (DN) and then that
resulting vector can be considered either contravariant or covariant since both spaces are then Cartesian.
This is the case with A = B x C under rotations in 3D space. On can think of the εabc as moving in this
limit from a tensor density of weight -1 to an ordinary tensor (ε is treated in Section D.4 below).

8. If vector V is a vector density of weight W, then the four expansions shown in (7.13.10) become

V= ΣnVn un with un • V = Vn
V= ΣnVn un with un • V = Vn
V= JW Σn V'n en with en • V = JW V'n
V= JW Σn V'n en with en • V = JW V'n . (D.2.9)

Proof: If V has weight W then, according to item 6 above, the four dot products shown on the right are
scalar densities of weight W (the basis vectors all have weight 0). We then evaluate the four dot products :

un • V = (un)i Vi = δniVi = Vn // see (7.18.3)

un • V = (un)i Vi = δniVn = Vn // see (7.18.3)

(en • V ) = JW (e'n • V' ) JW (e'n)iV'i = JW δniV'i = JW V'n // see (7.18.1)

(en • V ) = JW (e'n • V' ) JW (e'n)iV'i = JW δniV'i = JW V'n // see (7.18.1) (D.2.10)

255
 Appendix D: Tensor Densities

In the last two lines, we use item 6 that A•B = J+(W+w) A'•B' with w=0.
Thus the two un expansions are unaltered, but the en expansions pick up an extra factor of JW. For
example, assume that V = Σmαm em with αm unknown. Then

V • en = (Σmαm em)• en = Σmαm(em• en) = Σmαmδmn = αn

Therefore αn = V • en = JW V'n so the expansion must be V = JW Σm V'n em .

9. This same idea applies to the more general tensor expansions of Appendix E. For example, if Aijk is
a tensor density of weight W, one will have in place of (E.2.9),

A = JW Σijk A'ijk (ei⊗ej⊗ek) . (D.2.11)

Proof: According to item 6, the direct product space dot product shown below is a scalar density of
weight W, so

A • (ea⊗eb⊗ec) = JW A' • (e'a⊗e'b⊗e'c) = JW A'ijk (e'a)i (e'b)j (e'b)k = JW A'abc

If we write A = Σijk αijk (ei⊗ej⊗ek) with αijk unknown, we find as in the vector case above that the
coefficients are αijk = JW A'abc.

D.3 Theorem about Totally Antisymmetric Tensors: there is really only one: εabc...

The permutation tensor εabc...x has a number of indices equal to the dimension N of the space in which
one is working. For N = 3 one has εabc, and εab or εabcd do not exist. The normal calibration of the tensor
is ε123..N = 1. Each index swap causes a minus sign such as εabc = -εbac = - εcba , and if any pair of
indices are the same, the result is zero, such as ε131 = 0, Due to this swapping property on any index, this
tensor is called "totally antisymmetric".

Theorem: Apart from a scalar factor, there exists only one totally antisymmetric (TA) tensor. (D.3.1)

Proof: Suppose there were two TA tensors called εabc... and rabc.... If two or more of the indices are
equal, both tensors are 0, so for such index sets, one can say rabc.. = f εabc.. where f is any finite
function whatsoever. Consider now the case where all the indices are distinct and therefore exhaust the set
123...N, and consider abc... to be a permutation of 123...N obtained by doing S pairwise swaps,

abc... = P(123...) p = (-1)S . (D.3.2)

If one were to associate a sign change with each swap, the total sign change would be p, the parity. Since
ε and r are both TA tensors, each tensor can be "unwound" back to a standard index order by doing these
S swaps, and the swaps will cause a total sign of p relative to that standard order, so

rabc.. = p r123... // for example, r2134.. = (-1)1 r1234.. (D.3.3)

εabc.. = p ε123...

256
 Appendix D: Tensor Densities

Define scalar function f ≡ r123... / e123... , whatever it might be. Then

rabc.. = p(f ε123...)

εabc.. = p ε123... (D.3.4)

and dividing these two equations one finds,

rabc.. = f εabc.. (D.3.5)

which has now been shown valid for all index sets abc.. . Therefore, any "other" totally antisymmetric
tensor is just a scalar function times the ε tensor.

Alternate proof assuming (D.11.8) below. Assume rabc.. = f εabc... Apply εabc.. to both sides and sum
on abc.. to get εabc..rabc.. = εabc..(f εabc..) = (εabc.. εabc..)f = N! |g| f by (D.11.8). Therefore the
assumed form rabc.. = f εabc.. is valid with f = (N!|g|)-1εabc..rabc.. .

Similar Example: For N= 3, if rank-4 tensor Mabcd is totally antisymmetric on indices abc, then Mabcd
can be written in the form Mabcd = εabcqd where qd is a vector. In fact qd = (3!|g|)-1 εabcMabcd.

D.4 The contravariant ε tensor

The "Levi-Civita symbol" or tensor is one that equals the permutation tensor when all indices are "up", as
we shall see (a convention). One does not raise and lower indices of the mechanical permutation "tensor",
but one does do this on the Levi-Civita tensor, which we shall just call "the ε tensor". As we shall also
see, the ε tensor is really a tensor density of weight -1. When writing εabc.. one must have clearly in
mind whether one means the permutation tensor or the Levi-Civita tensor. Unless indices are all up, they
are in general not the same.

Knowing nothing to start, assume that the totally antisymmetric εabc.. tensor transforms under F as a
tensor density of some weight W which we hope to determine. Then

ε'abc.. = J-W Raa' Rbb' ... εa'b'c'.. . (D.4.1)

Assume that εabc.. is the usual permutation tensor normalized to ε123...N = +1. This is the convention
used by Weinberg p 99. This means each index swap changes the sign, and if two or more indices are the
same, ε = 0. This is an important starting assumption, and from it most everything follows.
Given this assumption, the RHS of (D.4.1) is totally antisymmetric (TA). The argument is given once
here and then used later several times. Consider an a↔b swap. Then

ε'bac.. = J-W Rba' Rab' ... εa'b'c'.. = J-W Rbb' Raa' ... εb'a'c'.. // b'↔a' dummy indices

= J-W Raa' Rbb' ... (–εa'b'c'..) = – ε'abc.. . (D.4.2)

257
 Appendix D: Tensor Densities

The same result is true for any swap, thus RHS (D.4.1) is TA. Since according to Section D.3 there is
only one TA tensor available, apart from a scalar function factor, it follows that

ε'abc.. = Kεabc... (D.4.3)

where K is some scalar function, perhaps just a constant. Equation (D.4.1) above then reads

K εabc... = J-W Raa' Rbb' ... εa'b'c'.. . (D.4.4)

Setting in the standard order, one finds that

K ε123... = J-W R1a' R2b' ... εa'b'c'.. (D.4.5)

or
K = J-W det(Rij) = J-W (J)-1 = J-(W+1) // J = det(S) = 1/det(R) (D.4.6)

so now

ε'abc.. = Kεabc... = J-(W+1) εabc... (D.4.7)

A second assumption is now made: that εabc.. (contravariant!) has the same value structure in any
frame of reference, which is to say it is the same in x'-space as it is in x-space,

ε'abc.. = εabc... . (D.4.8)

where εabc... is the usual permutation tensor with ε123... = +1. This assumption is consistent with
taking W = -1 in (D.4.7).

Fact: The Levi-Civita tensor εabc... has weight W = -1. From (D.2.5) this is true regardless of the index
positions on the ε tensor. (D.4.9)

Again, this follows the convention of Weinberg p 99. Some authors instead arrange for the above
equation to be true for the covariant ε tensors, and use then ε123...N = ε'123...N = +1, but we shall
follow Weinberg.
To summarize, assuming that εabc... is the usual permutation tensor normalized in the usual way,
and assuming that ε'abc.. = εabc... so this tensor is the same in all frames or spaces, THEN one
concludes that εabc... must transform as a rank-N tensor density of weight W = -1. That is to say,

ε'abc.. = J Raa' Rbb' ... εa'b'c'.. // this is (D.4.1) above with W = -1 (D.4.10)

Viewed in this light, the tensor εabc.. is known as the Levi-Civita tensor.

Tullio Levi-Civita (1873-1941). Italian, University of Padua 1892, with Ricci published the theory of
tensor algebra in 1900 (see Refs.), which work assisted Einstein circa 1915 in formulating the theory of
general relativity. The ε tensor bears his name. Sometimes the affine connection (Appendix F) is called
the Levi-Civita connection.

258
 Appendix D: Tensor Densities

D.5 Some facts about the ε tensor

1. Consider, with the assumption of (D.2.4) above applied to the ε tensor,

εabc... = gaa' gbb'..... εa'b'c'... // just as with any other tensor (D.5.1)

This is again in the convention of Weinberg p 99 (4.4.10). Although we show all indices lowered at once,
one can lower them one at a time in the usual manner using gij.

Added-sign-s convention. Some authors make a special exception for the ε tensor and introduce an
extra sign s into the above equation ( recall that s = -1 for special relativity)

εabc... = s gaa' gbb'..... εa'b'c'... s = sign[det(gij)] . (D.5.2)

This convention then invalidates the idea (7.4.11) that gab lowers a tensor index on ε and gab raises a
tensor index on ε. If that idea were valid, the above would read εabc.. = s εabc and would force s =1.
Furthermore, the meaning of a mixed ε tensor like εab becomes undefined in this convention because
we don't know how to obtain εab from say εab. Since this convention is nevertheless used by some
authors, we shall present some results below in this added-sign-s convention.

Up-Equals-Down convention. Recall that we use the Weinberg convention (D.5.1), which says ε's
indices work like those of any other tensor, and (D.4.8) which says ε'abc.. = εabc... . Some
authors instead replace the "fiat rule" ε'abc.. = εabc... by a different fiat rule : ^εabc... = εabc...
where we use a hat to distinguish their low index ε from our low index ε. Below we will show in the
Weinberg convention that εabc... = g εabc... so we can then identify ^εabc... = (1/g) εabc.... The
upper index version is the same, so ^εabc.. = εabc.. . In this convention one finds that ^εabc.. is a
tensor density of weight - 1 and then ^εabc... has weight 2 - 1 = +1. We mention this convention only
because it exists, but we shall not express any of our results in this "up equals down" convention.

Continuing with (D.5.1), install the reference sequence to obtain

ε123.. = g1a' g2b'..... εa'b'c'... = det(gij) = g . // see (D.2.2) (D.5.3)

Similarly, ε'123.. = det(g'ij). To summarize,

ε123.. = det(gij) = g // ε all-down index reference values

ε'123.. = det(g'ij) = g' . (D.5.4)

In the "added sign s convention", these last two equations would have sg = |g| and sg' = |g'| on the right
which means then these two ε values would be always positive.

259
 Appendix D: Tensor Densities

2. Take the same starting point as above

εabc... = gaa' gbb'..... εa'b'c'... (D.5.1)

The RHS is a totally antisymmetric in indices abc... . This is by argument (D.4.2) repeated in this context,

εbac... = gba' gab'..... εa'b'c'... = gbb' gaa'..... εb'a'c'...

= gbb' gaa'..... (-εa'b'c'...) = - gaa'gbb' ..... εa'b'c'... = - εabc... (D.5.5)

Then according to Section D.3, the right side of the above can be written as

RHS = C εabc... (D.5.6)

since we there is only one TA tensor apart from scalar C. Therefore

εabc... = C εabc... . (D.5.7)

Insert the reference sequence

ε123... = C ε123... = C (D.5.8)

But ε123... = det(gij) from (D.5.3), so

C = det(gij) = g (D.5.9)

and then (D.5.1) says, for the "Weinberg convention",

εabc... = det(gij) εabc... = g εabc... // relating all down to all up

ε'abc... = det(g'ij) ε'abc... = g' ε'abc... // = g' εabc... . (D.5.10)

where the second line follows by the same argument. These equations relate all indices down to all up in
the same space. Notice that both εabc... and ε'abc... are totally antisymmetric.

In the "added sign s convention" the above equations are instead

εabc... = |det(gij)| εabc... = |g| εabc... // relating all down to all up

ε'abc... = |det(g'ij)| ε'abc... = |g'| ε'abc... // = |g'| εabc... (D.5.11)

3. Divide the two Weinberg convention equations in (D.5.10) to find that

ε'abc... = [det(g'ij)/ det(gij)] εabc... = (g'/g) εabc... (D.5.12)

From (S.1.2) one has (g'/g) = J2 so the conclusions regarding ε are these:

ε'abc... = J2 εabc... = (g'/g) εabc.. ε'abc... = εabc... = permutation tensor // general

εabc... = permutation tensor if g=1 (D.5.13)

260
 Appendix D: Tensor Densities

These conclusions are valid for the "added sign s convention" as well since det(g) and det(g') always have
the same sign as shown in (D.1.2).

Two comments:

• Although we set ε'abc... = εabc... by fiat, we cannot similarly set ε'abc... = εabc...by fiat. This latter
result comes out being ε'abc... = J2 εabc... as just shown. (D.5.14)

• The fact that ε'abc... = J2 εabc... does not say that εabc is a tensor density of weight -2 because there
are no R factors showing (see tensor density definition (D.1.4) ) . (D.5.15)

D.6 The covariant ε tensor : repeat Section D.4 as if its weight were not known

According to (D.2.5), lowering indices does not change the weight of a tensor density. Eq (D.4.9) shows
that εabc.. is a tensor density of weight -1, so we know right away that εabc.. is also a tensor density of
weight -1. Nevertheless, it is interesting (and tests our consistency) to see what happens when the same
method used in Section D.4 for εabc... is applied to εabc... .
We start by assuming εabc.. is a tensor density of some unknown weight W,

ε'abc.. = J-W [ Raa' Rbb'.... εa'b'c'..] . (D.6.1)

Eq (D.5.10) shows that εa'b'c'.. is totally antisymmetric (TA) since we know εabc... is TA. The
argument of (D.4.2) with up↔down indices then shows from (D.6.1) that ε'abc... is TA as well. Since
according to Section D.3 there is only one TA tensor available, apart from a scalar function factor, it
follows that RHS(D.6.1) = K εabc.. so (D.6.1) then says,

K εabc.. = J-W [ Raa' Rbb'.... εa'b'c'..] . (D.6.2)

Use (D.5.10) to set εa'b'c'.. = det(gij) εa'b'c'.. inside the bracket,

K εabc.. = J-W [ Raa' Rbb'.... det(gij) εa'b'c'..] , (D.6.3)

and then install the reference sequence on both sides

K ε123.. = J-W [ R1a' R2b'.... det(gij) εa'b'c'..] . (D.6.4)

But (D.5.3) says ε123.. = det(gij), so cancel det(gij) on both sides to get [ recall Rij = Sji ]

K = J-W [ R1a' R2b'.... εa'b'c'..] = J-W det(Rij) = J-W det(Sji) = J-W J = J-(W-1) . (D.6.5)

So here the result is K = J-(W-1) whereas in (D.4.6) the result was K = J-(W+1) . In the current case, since
(D.6.1) and (D.6.2) have the same RHS, setting the LHS's equal says

261
 Appendix D: Tensor Densities

ε'abc.. = K εabc.. = J-(W-1) εabc.. (D.6.6)

But (D.5.13) says that ε'abc.. = J2 εabc.. and therefore W = -1.

Thr conclusion is that εabc... transforms with weight -1, the same as εabc..., so (D.6.1) becomes

ε'abc.. = J [ Raa' Rbb'.... εa'b'c'..] . (D.6.7)

D.7 Generalized cross products

In (A.4.1) and (A.4.2) the following cross product of N-1 vectors is considered (now in Standard
Notation, all indices are contracted except a)

Qa ≡ εabc...x BbCcDd.....Xx or Q = B x C x D .... x X . (D.7.1)

If the vectors B,C,D..X are all contravariant vectors, then applying the rule (D.2.3), one concludes that,
since ε is a tensor density of weight -1 and since all the RHS vectors have weight 0, the object Qa is a
covariant vector density of weight = -1, and thus has this transformation rule

Q'a = J RabQb . (D.7.2)

Similarly, one may consider

Qa ≡ εabc...x BbCcDd.....Xx . (D.7.3)

If vectors B,C,D...X are covariant vectors, then Qa is a vector density of weight -1 and

Q'a = J RabQb . (D.7.4)

D.8 The tensorial nature of curl B

It has just been shown that C = A x B [ Ca ≡ εabc AbBc ] is a vector density of weight -1, this being a
special case of the generalized cross product discussion above. As noted in the Example below (D.2.8), if
gij = 1 and r is a (global) rotation, then g'ij = 1 as well. This means from (D.1.2) that g = g' = 1 and J =
1. Then the weight -1 transformation rule C'a = J RabCb becomes C'a = RabCb which is the same as a
weight 0 rule, so then C can be regarded as a regular tensor in its transformation properties (in this special
case). [ Under inversion, if A→ -A and B→ -B, one gets C→ +C, so C is sometimes called a
pseudotensor. ]

One might conjecture that C = ∇ x B is also a vector density of weight -1, and that conjecture is correct
as is now shown. Consider

Cn = εnab ∂aBb (D.8.1)

262
 Appendix D: Tensor Densities

where B is assumed to be an tensorial vector. It is helpful to write this equation in the following manner

Cn = εnab [ ∂aBb – ∂bBa ]/2 (D.8.2)

where the second term is the same as the first term, since

– εnab ∂bBa = εnba ∂bBa = εnab ∂aBb . (D.8.3)

Recall now from (7.16.4) that the covariant derivative of a vector is given by

Bb;a = ∂aBb – Γcab Bc (D.8.4)

where the affine connection Γcab is symmetric under a↔b. Therefore,

Bb;a – Ba;b = [∂aBb – Γcab Bc] - [∂bBa – Γcba Bc] = ∂aBb – ∂bBa . (D.8.5)

Therefore Cn can be expressed as

Cn = εnab [Bb;a – Ba;b ]/2 (D.8.6)

so by the same ε anti-symmetry noted above the final result is

Cn = εnab Bb;a . (D.8.7)

The major feature of Bb;a -- as discussed in (7.16.5) -- is that it is a rank-2 tensor if B is a tensorial
vector. The weight addition rule (D.2.3) can then be applied to εnab Bb;a. Since εnab has weight -1 and
Bb;a has weight 0, the conclusion is that Cn is a vector density of weight -1.

Thus, C = ∇ x B = curl B is a vector density of weight -1.

D.9 Tensor E as a weight 0 version of ε : three conventions

1. Equations in the "Weinberg Convention"

In this Section it is assumed as in (D.2.4) that gij and gij raise and lower indices of the ε tensor just as
they do for any other tensor (Weinberg convention). Above it was established that

εabc... = g εabc... ε123.. = +1 ε123... = g (D.5.10)

ε'abc... = g'ε'abc... ε'123.. = +1 ε'123... = g' (D.5.10)

ε'abc... = J2 εabc... = (g'/g) εabc... (D.5.13)

ε is a rank-N tensor of weight W = -1 (D.4.9) and (D.6.7) (D.9.1)

263
 Appendix D: Tensor Densities

Again, just in passing, notice that ε'abc...= J2 εabc...does not say ε has weight -2 because the R factors
are not present on the right side.
Consider now the following new objects defined by (sg = |g|, s= sign(g) = sign(g') as in (D.1.2) ),

Eabc... ≡ |g|-1/2 εabc... ⇒ E123... = |g| -1/2 g = |g| -1/2 s |g| = s |g|1/2
E'abc... ≡ |g'|-1/2 ε'abc.. ⇒ E'123... = |g'| -1/2 g' = |g'| -1/2 s |g'| = s |g'|1/2 .
(D.9.2)

From (D.1.2) g' = J2g so that g transforms as a scalar density of weight -2. Since the sign of g and g' are
the same, if follows that (sg) = |g| is also a scalar density of weight -2, and then the
quantity |g|-1/2 transforms as a scalar density of weight +1, since |g'|-1/2 = J-1 |g|-1/2. Looking at the
equation Eabc... ≡ |g|-1/2 εabc... above, and using the weight summation rule (D.2.3), one concludes
that Eabc... transforms as a tensor of weight (+1) + (-1) = 0, and so Eabc... is an ordinary rank-N tensor.

Comment: It was shown near (7.15.9) that a tensor density equation with matching weights is "covariant",
so one is not surprised to see the second line of (D.9.2) having the same form as the first line but
everything is primed (s = s').

Raising indices on both sides of (D.9.2) gives these contravariant ordinary rank-N tensors,

Eabc... ≡ |g|-1/2 εabc... ⇒ E123... = |g|-1/2

E'abc... ≡ |g'|-1/2 ε'abc... ⇒ E'123... = |g'|-1/2 . (D.9.3)

To compare Eabc... and Eabc... we first evaluate Eabc... ,

Eabc... = |g|-1/2 εabc... // lower indices on both sides of 1st Eq of (D.9.3) since covariant
= |g|-1/2 g εabc... // from (D.9.1)
= s |g|-1/2 |g| εabc... // since g = s|g| by (D.1.2)
= s |g|1/2 εabc... (D.9.4)

We then pair this last equation with the first of (D.9.3),

Eabc... = s |g|1/2 εabc... (D.9.4)

Eabc... = |g|-1/2 εabc... (D.9.3) (D.9.5)

Dividing the two equations gives Eabc... / Eabc... = s |g| = g so that

Eabc... = g Eabc... (D.9.6)

We can then construct an E E product as follows

Eabc... EABC... = [ |g|-1/2 εabc... ] [ |g|-1/2 εABC...] // (D.9.3) and 1st line (D.9.4)

= |g|-1 εabc... εABC... . (D.9.7)

264
 Appendix D: Tensor Densities

Next, we use standard ordering to get

E123... = |g|-1/2 // from (D.9.3)

E123... = s|g|+1/2 . // from (D.9.4) (D.9.8)

Summarizing (D.9.8) and (D.9.6) and (D.9.7),

E123... = |g|-1/2 E123... = s|g|+1/2 Eabc... = g Eabc... = s|g| Eabc...

E'123... = |g'|-1/2 E'123... = s|g'|+1/2

Eabc... EABC... = |g|-1 εabc... εABC...

E'abc... E'ABC... = |g'|-1 ε'abc... ε'ABC... (D.9.9)

Since |g|-1 is a scalar density of weight +2 and each ε has weight -1 and each E has weight 0, one is
happy to see the weights balance of the two sides of this pair of covariant equations.

2. Equations in the "added sign s convention"

The previous section shows how things work out using the "Weinberg convention" noted at the start of
Section D.5. Here is the previous section redone in the "added sign s convention" where εabc... has an
extra sign s as shown in (D.5.2). To get from Weinberg to added-sign-s, take εabc... → sεabc.... Doing
this change in (D.5.10) for both εabc.. and ε'abc.. gives

εabc... = sgεabc... ε123.. = +1 ε123... = sg = |g| // g → sg

ε'abc... = sg'ε'abc... ε'123.. = +1 ε'123... = sg' = |g'| // g' → sg' (D.9.10)

ε'abc... = |J|2 εabc... = (g'/g) εabc... ε is rank-N tensor of weight W = -1 // same

Consider the following new objects defined by ( sg = |g|, s= sign(g) = sign(g') as in (D.1.2) )

Eabc... ≡ |g| -1/2 εabc... ⇒ E123... = |g| -1/2 |g| = |g|1/2

E'abc... ≡ |g'| -1/2 ε'abc... ⇒ E'123... = |g'|-1/2 |g'| = |g'|1/2 . (D.9.11)

By the same argument given above, Eabc... is an ordinary covariant tensor (ie, weight = 0). However, the
indices cannot be raised by gij. In this convention then one must make independent definitions of the
contravariant components as follows,

Eabc... ≡ |g|-1/2 εabc... ⇒ E123... = |g|-1/2

E'abc... ≡ |g'|-1/2 ε'abc... ⇒ E'123... = |g'|-1/2 (D.9.12)

To compare Eabc... and Eabc... ,

265
 Appendix D: Tensor Densities

Eabc... = |g|-1/2 εabc... = |g|-1/2 |g| εabc... = |g|-1/2 |g| εabc...

Eabc... = |g|-1/2 εabc... (D.9.13)

so that
Eabc... = |g| Eabc... (D.9.14)
Summarizing,

E123... = |g|-1/2 E123... = |g|+1/2 Eabc... = |g| Eabc...

E'123... = |g'|-1/2 E'123... = |g'|+1/2 E'abc... = |g'| E'abc...

Eabc... EABC... = |g|-1 εabc... εABC...

E'abc... E'ABC... = |g'|-1 ε'abc... ε'ABC... (D.9.15)

In this "added sign s" convention, all these summarized results involve only |g| and there are no factors of
s floating around. The cost of this benefit is a lack of true covariance (when s = -1), as demonstrated in
Section D.11 below.

3. Equations in the "Ricci-Levi-Civita convention"

Ricci and Levi-Civita use the "added s convention" but add a factor σ = sign(det(S)) into their definition
of E ( see their paper p 135 or Hermann pp 31-21) so that

Eabc... ≡ σ|g| -1/2 εabc... ⇒ E123... = σ|g| -1/2 |g| = σ|g|1/2

E'abc... ≡ σ|g'| -1/2 ε'abc... ⇒ E'123... = σ|g'|-1/2 |g'| = σ|g'|1/2

Eabc... ≡ σ|g|-1/2 εabc.. ⇒ E123... = σ|g|-1/2

E'abc... ≡ σ|g'|-1/2 ε'abc... ⇒ E'123... = σ|g'|-1/2 (D.9.16)

Summarizing,

E123... = σ|g|-1/2 E123... = σ|g|+1/2 Eabc... = |g| Eabc...

E'123... = σ|g'|-1/2 E'123... = σ|g'|+1/2 E'abc... = |g'| E'abc...

Eabc... EABC... = |g|-1 εabc... εABC...

E'abc... E'ABC... = |g'|-1 ε'abc... ε'ABC... (D.9.17)

Notice that in all three conventions, the last equation pair is the same.

Since Ricci and Levi-Civita did not raise and lower individual indices in their 1900 paper, they were not
concerned about their convention being non-covariant in that sense.

266
 Appendix D: Tensor Densities

D.10 Representation of ε, εε and contracted εε as determinants

This section is in developmental notation and ε is the permutation tensor.

1. Theorem about a certain permutation sum

Consider the following object Q defined as a signed permutation sum of the product of N matrix elements
of a matrix Mij,

Qabc..x ≡ ΣP p P2(Ma1Mb2 Mc3.....MxN) (D.10.1)

In this equation, P2 represents a permutation of the set of 2nd indices of the N matrix elements, and the
sum is over all N! such permutations.
There are many ways to arrive at a given permutation of 123...N by doing pairwise swaps, but for all
these ways, the number of swaps S will be either even or odd. The parity p of a permutation is defined
then as (-1)S and this p appears in the above sum.
If one were to swap indices 2↔3 on the right above, each permutation would have S→ S+1 since an
extra swap is needed to undo 2↔3. Thus, all parities p → -p and in fact the whole object negates. But the
swap 2↔3 is the same as b↔c since Mb3 Mc2 = Mc2 Mb3. Applying this argument to any pair of
indices, one concludes that Qabc..x is totally antisymmetric and therefore by (D.3.1) can be written as K
εabc...x :

ΣP p P2(Ma1Mb2 Mc3.....MxN) = K εabc...x . (D.10.2)

where in this Section εabc...x is just the permutation tensor (not the Levi-Civita tensor).

Setting abc..x to 123..N, one gets.

ΣP p P2(M11M22 M33.....MNN) = K . (D.10.3)

The left side of this last equation can be written as

ΣP p P2(M11M22 M33.....MNN) = Σabc..x εabc...x M1aM2b M3c.....MNx (D.10.4)

because p = εabc...x correctly assesses the parity of any given permutation. But this object is simply
det(M) so the conclusion is that K = det(M) and then

ΣP p P2(Ma1Mb2 Mc3.....MxN) = det(M) εabc...x . (D.10.5)

Consider now the following matrix where abc..x is some permutation of 123...x,

M(abc..) = Ma1 Ma2 Ma3 ... MaN

Mb1 Mb2 Mb3 ... MbN
Mc1 Mc2 Mc3 ... McN
...
Mx1 Mx2 Mx3 ... MxN . (D.10.6)

267
 Appendix D: Tensor Densities

By rearranging the rows into their normal numerical order, one obtains matrix M, but incurs a sign from
the various row swaps which sign is just εabc..x. Therefore,

det(M(abc..) ) = εabc...x det(M) (D.10.7)

and therefore we obtain the following "theorem" :

ΣP p P2(Ma1Mb2 Mc3.....MxN) = det(M(abc..) ) = det(M) εabc...x . (D.10.8)

The permutation sum is thus just the determinant of matrix M(abc..). The first term in the permutation
sum, the term with an identity permutation, corresponds to the product of the diagonals of M(abc..).

2. Application of the theorem to M = δ: a representation of ε

Apply the above theorem to matrix M = 1 ≡ δ, the identity matrix, so Mij = δi,j. Clearly det(δ) = 1 and
one then has

ΣP p P2(δa,1δb,2 δc,3.....δx,N) = det[δ(abc..) ] = εabc...x . (D.10.9)

Thus is obtained a well-known representation of εabc..x as a certain determinant of Kronecker deltas,

εabc...x = det[δ(abc..) ]

where δ(abc..) = δa,1 δa,2 δa,3 ... δa,N = Ra

δb,1 δb,2 δb,3 ... δb,N = Rb
δc,1 δc,2 δc,3 ... δc,N = Rc
...
δx,1 δx,2 δx,3 ... δx,N = Rx (D.10.10)

For future use, each row vector has been given a name like Ra where (Ra)i = δa,i .

The conclusion then is that (vertical bars here mean determinant) :

(D.10.11)
which is the same as

εabc...x = ΣP p P2(δa,1δb,2 δc,3.....δx,N) . (D.10.12)

268
 Appendix D: Tensor Densities

3. Outer product of two ε tensors.

Consider now

⎛ RRab ⎞ ⎛ RRa' ⎞
... ⎟ ... ⎟
= det ⎜ = det ⎜
b'
εabc...x and εa'b'c'...x' . (D.10.13)
⎝ Rx ⎠ ⎝ Rx' ⎠
Then

⎛ RRab ⎞ ⎛ RRa' ⎞ ⎛ RRab ⎞

... ⎟ ⎜ ... ⎟ ... ⎟
εabc...x εa'b'c'...x' = det ⎜ = det ⎜
b'
det det ( Ra' Rb' ... Rx')
⎝ Rx ⎠ ⎝ Rx' ⎠ ⎝ Rx ⎠

⎛ RRab ⎞
... ⎟ a' b'
= det { ⎜ ( R R ... Rx') } (D.10.14)
⎝ Rx ⎠
which is the determinant of this matrix

Ra• Ra' Ra• Rb' Ra• Rc' ... Ra• Rx'

Rb• Ra' Rb• Rb' Rb• Rc' ... Rb• Rx'
Rc• Ra' Rc• Rb' Rc• Rc' ... Rc• Rx'
...
Rx• Ra' Rx• Rb' Rx• Rc' ... Rx• Rx' . (D.10.15)

A typical element of this matrix is given by

Rc• Rb' = (Rc)i(Rb')i = δc,i δb',i = δc,b' (D.10.16)

so that matrix can be written as

δa,a' δa,b' δa,c' .... δa,x'

δb,a' δb,b' δb,c' .... δb,x'
δc,a' δc,b' δc,c' .... δc,x' ≡ δ(abc..x; a'b'c'..x')

....
δx,a' δx,b' δx,c' .... δx,x' (D.10.17)

where we have made up a name for this matrix as shown.

The conclusion then is that

269
 Appendix D: Tensor Densities

(D.10.18)
which is the same as

εabc...x εa'b'c'...x' = ΣP p P2(δa,a'δb,b'δc,c'.....δx,x') (D.10.19)

because the right side of (D.10.19) is precisely the determinant appearing in (D.10.18). As usual, the
argument of P2 is the product of the diagonal elements of the matrix of interest.

4. Contracting the first index of the outer product of two ε tensors.

Consider what happens if one sums on the first index of the εε product:

Σa εabc...x εab'c'...x' . (D.10.20)

For fixed given values of bc..x and b'c'...x' , there is only one way this sum can be non-zero. In that one
way, bc..x and b'c'...x' must each be permutations of the set {12..N exclude A} where A is the "hit value"
of a in the sum on a. In the sum, only a = A contributes. Then using this hit value A and (D.10.19),

Σa εabc...x εab'c'...x' = εAbc...x εAb'c'...x'

= ΣP p P2(δA,Aδb,b'δc,c'.....δx,x') = ΣP p P2(δb,b'δc,c'.....δx,x') (D.10.21)

where in this last expression the sum can be regarded as being over permutations where b'c'...x' is a
permutation of b,c..x. Each of these lists of integers is in turn a permutation of {12..N exclude A}. Now,
parity p = (-1)S where S is a number of swaps it takes to connect b'c'...x' with b,c..x, since a = a' = A. One
might wonder if the overall sign of the RHS of the last equation is correct. A check of the first term in this
sum which is just δb,b'δc,c'.....δx,x' shows that this overall sign is indeed correct. This first term must
be positive because the product of two ε's is either +1 or 0. As an example,

Σa εabc εab'c' = ΣP p P2(δb,b'δc,c') = δb,b'δc,c' – δb,c'δc,b' . (D.10.22)

The permutation sum shown on the right side of (D.10.21) is the determinant of δ(abc..x; a'b'c'..x')
but with the first row and column crossed out. It can then be thought of as either the minor or cofactor of
the element aa of this big δ matrix. Therefore,

Σa εabc...x εab'c'...x' = [cof δ(abc..x; a'b'c'..x')

]aa (D.10.23)

270
 Appendix D: Tensor Densities

where the notation cofM refers to a matrix of cofactors with elements [cofM]ij. Don't confuse the a on
the right side with the local dummy summation index a on the left side.

The conclusion then is that (implied summation on a on the LHS)

(D.10.24)

5. Contracting two or more indices of the outer product of two ε tensors.

Consider what happens if one sums on the first two indices of the εε product:

Σa,b εabcd...x εabc'd'...x' . (D.10.25)

For fixed given values of c,d..x and c'd'...x' , in order for this double sum to be non-zero, the index sets
cd..x and c'd'...x' must each be permutations of the set {12..N exclude A,B} where A,B are a pair of hit
values for the a and b sums. If a=A and b=B is a hit value, then so is a=B and a=A, so there are 2!
contributing terms in the sum, and each term is +1. Therefore

Σa,b εabcd...x εabc'd'...x' = 2! εABc...x εABc'...x'

= 2! ΣP p P2(δA,AδB,B'δc,c'δd,d'.....δx,x') = 2!ΣP p P2(δc,c' δd,d'.....δx,x') (D.10.26)

where in this last expression the sum is over permutations where c'd'...x' is a permutation of c,d....x. Each
of these lists of integers is in turn a permutation of {12..N exclude A,B}. Now parity p = (-1)S where S is
a number of swaps it takes to connect c'd'...x' with c,d....x. Since the product of two ε's is either +1 or 0,
the overall sign of the right side shown must be correct. As an example,

Σa,b εabc εabc' = 2!ΣP p P2(δc,c') = 2 δc,c' . (D.10.27)

If c = c' = 2, then this says

Σa,b εab2 εab2 = ε132 ε132 + ε312 ε312 = 1 + 1 = 2 . (D.10.28)

The permutation sum shown on the right side of (D.10.26) is the determinant of δ(abc..x; a'b'c'..x')

but with the first 2 rows and columns crossed out. Therefore,

Σa,b εabcd...x εabc'd'...x' = 2! { [cof δ(abc..x; a'b'c'..x')

]aa}bb . (D.10.29)

The conclusion then is that (implied summation on a,b on the LHS)

271
 Appendix D: Tensor Densities

(D.10.30)

Here a and b on the left are just dummy summation indices, whereas a and b on the right indicate that the
right side is 2! {...} where {...} is the cofactor of a matrix which is the full δ(abc..x; a'b'c'..x') matrix
but with the first two rows and columns crossed out. One can imagine doing this crossing out in the
picture of the full δ matrix shown in (D.10.18). The diagonal element of the first row and column is δa,a'
and of the second row and column δb,b' so we use the notation aa and bb to denote these rows and
columns which are crossed out.

This pattern continues as more indices are contracted. If three indices a,b,c are contracted, there will then
be 3! hit values which are A,B,C and its permutations, and one just repeats the above discussion. The
result will then be

Σa,b,c εabcd...x εabcd'...x' = 3! {{ [cof δ(abc..x; a'b'c'..x')

]aa}bb}cc . (D.10.31)

The conclusion then is that (implied summation on a,b,c on the LHS)

(D.10.32)

Eventually one arrives at a point where all but one of the indices are summed, so that

εabcd...x εabcd...x' = (N-1)! |δx,x'| = (N-1)! δx,x' (D.10.33)

an example being

εabc2 εabc2 = 3! δ22 = 3! = ε1342 ε1342 + ε1432 ε1432 + 4 more terms = 1+1+4 = 6 . (D.10.34)

The final point is that at which all indices are summed, with result

εabcd...x εabcd...x = N! (D.10.35)

and example of which is

εabcεabc = ε1232 + ε2132 + 4 more terms = 1 + 1 + 4 = 6 . (D.10.36)

272
 Appendix D: Tensor Densities

6. Summary of Results

••••

εabcd...x εabcd...x' = (N-1)! δx,x'

εabcd...x εabcd...x = N! (D.10.37)

The determinants appearing above are sometimes called generalized Kronecker deltas. For example,
here is a sample determinant followed by our (easy to type) notation and then the official Kronecker delta
notation:

c d x
≡ δ(cd...x; c',d'...x')
≡ δ c' d' ... x' (D.10.38)

c
Often the deltas are written in the form δc,c' = δcc' or δc'

Using both our notation and the generalized Kronecker delta notation, we rewrite (D.10.37) this way

273
 Appendix D: Tensor Densities

a b c x
εabc..xεa'b'c'..x' = δ(abc...x; a'b'c'...x')
= δ a' b' c' .... x'

b c x
εabc..xεab'c'..x' = δ(bcd...x; b'c'd'...x')
= δ b' c' .... x'

c d x
εabcd..xεabc'd..x' = 2! δ(cd...x; c'd'...x')
= 2! δ c' d'.... x'

d e x
εabcde..xεabcd'e'..x' = 3! δ(de...x; d'e'...x')
= 3! δ d' e'.... x'
•••••
x
εabcd...x εabcd...x' = (N-1)!δ(x,x') = (N-1)!δx' = (N-1)! δx,x'

εabcd...x εabcd...x = N! (N = dimension of x-space) (D.10.39)

Keep in mind that in all of Section D.10, ε has represented the permutation tensor.

D.11 Covariant forms of the previous Section results

The Section D.11 results were in effect all developed in Cartesian x-space where up and down indices on
the ε's did not matter and we regarded ε as the permutation tensor. The rules for converting any of these
results to covariant form are as follows: [ after conversion, ε becomes the Levi-Civita tensor where
εab... is the only index position where it is the same as the permutation tensor ]

Weinberg convention: (D.11.1)

• Make the replacement ε***** ε***** → |g|-1 ε***** ε***** = E***** E***** from (D.9.9). In this
way we "continue off" the Cartesian equation (where g = 1) to obtain a covariant form. The objects
E***** and E***** are true contravariant and covariant (weight = 0) rank-N tensor. This is vaguely
reminiscent of "analytically continuing" a real function f(x) "off" the real axis to get a function of a
complex variable f(z). When evaluated on the real axis, f(z) and f(x) are the same. See the tensorization
discussion in Section 15.2 for more detail.

• Write the right side of the expression of interest replacing every δa,b = δab = gab as shown in (7.4.19).
Then the right side will also be a true tensor.

Example 1: For g = 1, consider the εε product with no summed indices [first line of (D.10.37)] for N = 2:

εabεa'b' =
⎪ δaa' δab' ⎪ = δ
a,a' δb,b' – δa,b' δb,a' . // ε = permutation tensor (D.11.2)
⎪ δba' δbb' ⎪

The covariant form is as follows, where now g is some arbitrary metric tensor for x-space,

274
 Appendix D: Tensor Densities

a a
EabEa'b' = |g|-1 εabεa'b' =
⎪ g ba' gbb' ⎪ = ga gb – ga gb
a' . (D.11.3)
⎪ g a' g b' ⎪ a' b' b'

The equation in x'-space would then be

a a
E'abE'a'b' = |g'|-1 ε'abε'a'b' =
⎪ g'a'b g'a'b' ⎪ = g'a g'a – g'a' g'a' (D.11.4)
⎪ g' b g' b' ⎪ b b' b b'

because true tensor equations are "covariant" as discussed in Section 7.15. One can raise and lower
individual indices to get for example these valid tensor equations which are 3 members of the family of 4!
= 24 tensor equations obtained by raising and lowering indices :

a a
ab
E Ea'b' = |g| -1 ab ⎪ g a' g b' ⎪ = ga gb – ga gb
ε εa'b' = gb gb (D.11.5)
⎪ a' b' ⎪ a' b' b' a'

a a
EabEa'b' = |g|-1 εabεa'b' =
⎪ g a' g b' ⎪ = ga g – ga g (D.11.6)
⎪ gba' gbb' ⎪ a' bb' b' ba'

EabEa'b' = |g|-1 εabεa'b' =

⎪ gaa' gab' ⎪ = g g – g g (D.11.7)
⎪ gba' gbb' ⎪ aa' bb' ab' ba'

and of course in x'-space the equations are the same but everything is primed.

For the Levi-Civita ε tensor, we then have this covariant version of (D.10.39) :

Eabc..xEa'b'c'..x' = |g|-1εabc..xεa'b'c'..x' = g(abc...x; a'b'c'...x')

Eabc..xEab'c'..x' = |g|-1εabc..xεab'c'..x' = g(bcd...x; b'c'd'...x')

Eabcd..xEabc'd..x' = |g|-1εabcd..xεabc'd..x' = 2! g(cd...x; c'd'...x')

Eabcde..xEabcd'e'..x' = |g|-1εabcde..xεabcd'e'..x' = 3! g(de...x; d'e'...x')
•••••
Eabcd...x Eabcd...x' = |g|-1εabcd...x εabcd...x' = (N-1)!g(x,x') = (N-1)!gxx

Ebcd...x Eabcd...x = |g|-1εabcd...x εabcd...x = N! (D.11.8)

where g(...) is the determinant of a matrix of gαβ elements. Eq. (D.11.5) provides an example of the first
line of (D.11.8) for the N=2 dimensions. Each equation in (D.11.8) is a true tensor equation allowing
individual non-summed indices to be taken up and down on both sides. And of course the contraction tilts
can be individually reversed as well. If all index positions are reversed, one gets a version of (D.11.8)
where all matrix entries have the form gαβ instead of gαβ, For example, the second last line would be,

Eabcd...xEabcd...x' = |g|-1εabcd...xεabcd...x' = (N-1)!g(x,x') = (N-1)!gxx' (D.11.9)

As in (7.4.19) one always has gαβ = gαβ = δαβ = δαβ = δα,β but of course gαβ ≠ δα,β .

275
 Appendix D: Tensor Densities

Since the above equations are true tensor equations, Section 7.15 says the equations are also valid in
x'-space where every object in the equation is primed.

Example 2: For N = 3 the second line of (D.11.8) reads,

b b
Eabc -1 abc
Eab'c' = |g| ε εab'c' = g( bc; ⎪ g cb' gcc' ⎪ = gb gc - gc gb . (D.11.10)
b'c') =
⎪ g b' g c' ⎪ b' c' b' c'

Raising b' and c' gives

bb' bc'
EabcEab'c' = |g|-1εabcεab'c' = g(bc; b'c'
) =
⎪ g cb' gcc' ⎪ = gbb'gcc' - gcb'gbc' .
⎪g g ⎪

Shuffling indices, this can be written

AA' AB'
EsABEsA'B' = |g|-1εsABεsA'B' = g(AB; A'B'
) =
⎪ g BA' gBB' ⎪ = gAA'gBB' - gBA'gAB'
⎪g g ⎪

= EsABEsA'B' = |g|-1εsABεsA'B' .
(D.11.11)

Added-sign-s and Ricci-Levi-Civita conventions:

Do the above two bullet items, then add an overall sign s to the right side, because
εabc.. = s g εabc... in these conventions instead of εabc.. = g εabc... so that
εabc..(Weinberg) = sεabc..(added-sign).

Example: The example above becomes ( εa'b'→ s εa'b')

a a
EabEa'b' = |g|-1 εabεa'b' = s
⎪ g ba' gbb' ⎪ = s ( ga gb – ga gb ) . (D.11.12)
⎪ g a' g b' ⎪ a' b' b' a'

The second equation is undefined (when s=-1), because we don't know how to lower just one index on ε,
hence on E, as discussed near (D.5.2). The third equation is

EabEa'b' = |g|-1 εabεa'b' =

⎪ gaa' gab' ⎪ = g g – g g
ab' ba' . (D.11.13)
⎪ gba' gbb' ⎪ aa' bb'

The first and third equations are true tensor equations, except individual indices cannot be raised and
lowered. If one were doing some significant work involving covariance and s=-1, it would certainly seem
advisable to use the Weinberg convention since it is completely "covariant" for either sign of s.

276
 Appendix D: Tensor Densities

D.12 How determinants of rank-2 tensors transform

In this document we have encountered only a few determinants of tensors (like gij) and tensor-like
objects (like Rij and Sij). Nevertheless, we would like to know how the determinant of a rank-2 tensor
transforms under x = F(x). To this end, we first rewrite the traditional determinant formula in a covariant
form. Once this is done, the conclusions come quickly.

Comment: The objects below like det(Mij) and det(Mij) of course have no dependence on the "indices" i
and j other than on the up/down position of these indices. Elsewhere we write these objects as det(M**)
and det(M**) where the "wildcards" just show the nature of the matrix elements (down-tilt or all-down).
Here, for technical reasons to be seen below, we maintain the notations det(Mij) and det(Mij).

We start with this mechanical statement of the determinant of a matrix Mij

det(Mij) = εab..x M1aM2b....MNx (D.12.1)

where ε is the permutation tensor. This form is "mechanical" just in the sense that if you drew a picture of
the matrix Mij and mechanically evaluated det(Mij), you obtain the above result. We can now restate
this determinant, switching from the permutation tensor εabc...x to the index-all-up Levi-Civita tensor.
As stated below (D.4.1), we take εabc..x to be equal to the permutation tensor. We then have

det(Mij) = εab..x M1aM2b....MNx (D.12.2)

which certainly looks more "covariant" since all summed indices appear to be contracted. But with fixed
upper indices 1,2...N "hanging out", it is not quite clear what is going on here. Since a through x are all
contracted, and since weight(ε) = - 1, one might be tempted to say det(Mij) is a scalar density of weight
-1, but that is incorrect.

To get a better view of things, it is useful to rewrite the above determinant as follows (proof follows)

det(Mij) = (1/N!) εAB...X εab..x MAa MBb ...MXx . (D.12.3)

Again, both ε's shown here are in effect permutation tensors.

We shall now show that the right sides of the previous two equations are exactly the same. First write the
right hand side of (D.12.3) as

RHS (D.12.3) = (1/N!) εAB...X { εab..xMAa MBb ...MXx } . (D.12.4)

The bracketed quantity can be expanded as,

QAB..X ≡ {εab..xMAa MBb ...MXx} = MA1 MB2 ...MXN + all signed permutations . (D.12.5)

meaning all signed permutations of the upper indices. Notice that

277
 Appendix D: Tensor Densities

Q12..N ≡ {εab..xM1a M2b ...MNx} = M11 M22 ...MNN + all signed permutations . (D.12.6)

This is just a mechanical evaluation of det(Mij) and we conclude that

Q12..N = det(Mij) . (D.12.7)

The tensor QAB..X is totally antisymmetric as this example demonstrates,

QBA..X = εabc..xMBaMAbMCc ...MXx = [ - εbac..x] MAbMBaMCc ...MXx

= - [ εabc..x] MAaMBbMCc ...MXx // a↔b

= - QAB..X . (D.12.8)

Then using theorem (D.3.1) we can write QAB..X in this form.

QAB..X = K εAB..X . (D.12.9)

To evaluate K, use the standard ordering 123..N to find from the above and (D.12.7),

Q12..N = K ε12..N = K = det(Mij) . (D.12.10)

Then, since K = det(Mij) one finds that

QAB..X = det(Mij) εAB..X . (D.12.11)

Therefore (D.12.4) can be written

RHS (D.12.3) = (1/N!)εAB...X { εab..xMAa MBb ...MXx }

= (1/N!)εAB...X {QAB..X }

= (1/N!)εAB...X {det(Mij) εAB..X}

= det(Mij) (1/N!) ΣAB..X (εAB...X)2 . (D.12.12)

The sum ΣAB..X (εAB...X)2 is a "sum of ones" and there is a one for each permutation of AB..X. All other
terms are zero. There are N! total permutations including the first, so

ΣAB..X (εAB...X)2 = N! (D.12.13)

which agrees with the last equation of (D.10.37). Therefore,

RHS (D.12.3) = det(Mij) (1/N!) ΣAB..X (εAB...X)2 = det(Mij) . (D.12.14)

278
 Appendix D: Tensor Densities

This concludes our too-lengthy proof that the right sides of (D.12.2) and (D.12.3) are identical.

Now we start with the proven result,

det(Mij) = (1/N!) εab..x εAB...X MAa MBb ...MXx . (D.12.3)

Eq (D.5.10) says that εABC... = g εABC... so the above can be written

det(Mij) = (1/g) (1/N!) εab..x εAB...X MAa MBb ...MXx . (D.12.15)

Now, finally, we have a form in which all tensor indices are contracted with no loose ends. We can then
use theorem (D.2.3) about the additivity of weights. Recall from (D.4.9) and (D.6.7) that both ε tensors
shown in (D.12.15) have weight -1, and that the object 1/g has weight +2 from (D.1.7). Adding, we find
that the object det(Mij) has weight 0.
We have therefore proven: (scalar = scalar density of weight 0)

Theorem: The determinant det(Mij) of a mixed rank-2 tensor Mij transforms as a scalar under the
transformation x = F(x). (D.12.16)

Comment: Since our Chapter 2 S matrix Sij is not a tensor, J = det(Sij) is not a scalar, and is fact not a
tensor of any kind since it bridges x-space and x'-space.

It is now straightforward to determine the transformation nature of the other three rank-2 tensor types, and
in fact to find simple relations between the four determinants. In all these cases, we use of the up-down
altering property of the g tensor of (7.4.11), the "up-tilt" or "down-tilt" version of matrix multiplication of
(7.8.11), the matrix rule det(AB) = det(A)det(B), facts (7.5.20) and (7.5.21), and the weight of g and 1/g
as determined in (D.1.6) and (D.1.7) :

det(Mij) = det(giaMaj) = det(gij) det(Mij) = g det(Mij).

-2 -2 0

det(Mij) = det(Miagaj) = det(Mij)det(gij) = { g det(Mij)}{g-1} = det(Mij)

0 -2 0 +2 0

det(Mij) = det(giaMaj) = det(gij) det(Mij) = g-1 det(Mij) = g-1 det(Mij) . (D.12.17)

+2 +2 0 +2 0

The weights are shown under each line of equations. The conclusions are these:

det(Mij) = det(Mij) // scalar densities of weight 0 ( = scalar)

det(Mij) = g det(Mij) // scalar density of weight -2 g = det(gij)

det(Mij) = g-1 det(Mij) // scalar density of weight +2 g-1 = det(gij) . (D.12.18)

279
 Appendix D: Tensor Densities

Here we have related each of the three partner determinants to the down-tilt determinant, and have shown
the weight of each type of determinant.

Example: We know that gij is a rank-2 tensor from Section 5.7. Thus, we expect to find from (D.12.18)
that,

det(gij) = det(gij) ⇒ det(δij) = det(δij) or 1 = 1 ok, weight 0

det(gij) = g det(gij) ⇒ g = g det(δij) = g*1 = g ok, weight -2

det(gij) = g-1 det(gij) ⇒ g-1 = g-1det(δij) = g-1* 1 = g-1 ok, weight 2 . (D.12.19)

Equations (D.12.18) are valid only for rank-2 tensors. They are not valid for R and S.

Go back now to the first line of (D.12.18),

det(Mij) = det(Mij) = det([MT]ji)

where MT is the "covariant transpose" of M as shown in (7.9.3). This shows that

det(M**) = det([MT]**)
or
det(Mut) = det(MTut) . // ut = up-tilt

Since det(M) = det(MT) is valid for any matrix regardless of index position (MT is the "matrix transpose"
of M), we find that

det(MTut) = det(MTut) = det(Mut) .

A similar argument beginning with

det(Mij) = det(Mij) = det([MT]ji) ⇒ det(Mdt) = det(MTdt)

shows that

det(MTdt) = det(MTdt) = det(Mdt) . // dt = down-tilt

Since MT = MT for both-up or both-down index positions, we arrive at this interesting generalization of
the traditional determinant theorem det(M) = det(MT) :

Fact: det(M) = det(MT) = det(MT) for all four possible index positions (D.12.20)

where T is the matrix transpose and T is the covariant transpose (7.9.3).

The scalar density weights of these determinants are given in (D.12.18) and do depend on the index
positions.

280
 Appendix E: Tensor Expansions

Appendix E: Tensor Expansions: direct product, polyadic and operator notation

E.1 Direct Product Notation

This entire Section uses the general Picture A context where x-space need not be Cartesian,

(E.1.1)

The Standard Notation of Chapter 7 is used throughout. Useful forms can be found in Section 7.18.

The key tool required for the expression of tensor expansions is the notion of a direct product (tensor
product, outer product) of n tensorial vectors defined in this simple way,

(A⊗B⊗C ...)abc... ≡ AaBbCc.....

(A⊗B⊗C ...)abc... ≡ AaBbCc..... etc . (E.1.2)

The tensor A⊗B⊗C ... is nothing more than the outer product of vectors A,B,C as in (7.1.1) for
contravariant vectors, but later extended to any mixture of vector types. One can, however, construct
outer products from more complicated tensors, see Lucht Tensor Products. Suppose one has

A = vector F,G = rank-2 tensors M,N = rank-3 tensors . (E.1.3)

Then here is a small sampling of outer products of tensors that can be formed,

(F⊗A)abc = FabAc (F⊗A⊗G)abcde = FabAcGde (F⊗G)abcd = FabGcd

(F⊗M⊗A)abcdef = FabMcdeAf (M⊗N)abcdef = MabcNdef . (E.1.4)

In any of these equations, any one or more of the indices can be lowered on both sides to provide a valid
equation since all objects in the list (E.1.3) are assumed to be true tensors. All the left-hand-side objects
above are tensors because the right sides show that they transform as tensors.

In what follows, only the (E.1.2) direct product of vectors shall be considered. One can define the dot
product of two direct-product-space vectors in this obvious manner,

(A⊗B⊗C ...) • (A'⊗B'⊗C' ...) ≡ (A⊗B⊗C ...)abc... (A'⊗B'⊗C' ...)abc

= AaBbCc..... A'aB'bC'c..... = A•A' B•B' C•C' ... (E.1.5)

where of course the indices abc can be "tilted" in any way desired according to (7.11.3).

281
 Appendix E: Tensor Expansions

E.2 Tensor Expansions and Bases

Preamble: A seeming paradox and how it is resolved

Before attacking tensor expansions below, we wish to head off a possible confusion between a quantity
transforming as a scalar under some transformation x' = F(x) versus transforming as a component of a
tensor. The "paradox" is presented in a few simple examples.

Example 1: Consider the vector r = (x,y,z) and a transformation r' = F(r) = Rr which is a rotation.
Assume x-space is Cartesian so g = 1. Consider the dot product s = r • ^ x . As noted in (5.10.2), this dot
product transforms as a scalar under F. In x'-space we still find that r' • ^
x ' = s. After all, the projection of
one vector onto another cannot change when the two vectors are rotated together.
On the other hand, direct calculation shows that r • ^ x = x, and x is a component of the vector r =
(x,y,z). So how can one say that quantity r • x is a scalar when r • ^
^ x = x which is a component of a
vector? This type of question arises frequently in the study of tensor analysis: there seems to be a paradox
that needs resolving.
The physicist considers several different "thought experiments".
In Experiment 1, two measurements are made. The first is made in Frame S and one finds that s = x.
The second measurement is made of rotated vector r' = Rr in rotated Frame S' and there one finds that s =
x'. The two measurements give the same number, so in Experiment 1 one finds that x' = x and the quantity
so measured is a scalar. In this experiment x = r • ^x and x' = r' • ^
x '.
In Experiment 2, everything is done in Frame S. One draws a vector r and measures x. One then
rotates this vector within Frame S to get a new vector r' = Rr where r' = (x',y',z') . One then measures x'
and finds x' ≠ x. In this experiment, x = r • ^
x and x' = r' • ^
x where ^x has no prime.
In Experiment 3 one observes unrotated r from rotated Frame S' so x = r • ^ x and x' = r • ^x ' and this
again results in x' ≠ x
Thus, in these three experiments, x has the same definition, but x' has three different definitions. That
is why one can have x' = x in Experiment 1 and x' ≠ x in Experiments 2 and 3. The concept of r • ^ x as a
^ ^
scalar applies to Experiment 1 where s = r • x = x = r' • x ' = x'.

Example 2. Consider the vector expansion (7.13.10) which says V = Σn V'n en with V'n = V • en . Here
we have the same paradoxical situation: we know that s = V • en must be a scalar with respect to F, yet it
is equal to the component of a contravariant vector V'n under F. If we write V = ΣnVnen, then in Frame S
we find that V • en = Vn which is like r • ^
x = x of the previous Example. In Frame S' we find instead that
V' • e'n = V'i(e'n)i = V'iδni = V'n which is like r' • ^
x ' = x' in Example 1. In Experiment 1 outlined
n n n
above, it is not a paradox to have s = e • V = V = V' be a scalar.

Example 3. Below we expand a rank-3 tensor A = Σijk αijk (bi⊗bj⊗bk) and we find that
αijk = A • (bi⊗bj⊗bk) where • is a the dot product (E.1.5). We know that this quantity αijk, analogous
to s in the two Examples above, must be a scalar under x' = F(x). Yet when we take bn = en we find in
(E.2.8) that αijk = A'ijk so our scalar αijk is equal to the component of a rank-3 tensor in x'-space.
Once again, in Experiment 1 this is not a contradiction. It just happens that in x'-space the scalar αijk

282
 Appendix E: Tensor Expansions

appears as the value of a tensor component A'ijk in x'-space, just the way s in Example 1 appears as the
value x' of vector r' in Frame S'.

Tensor Expansions

In what follows, we use the example of a rank-3 tensor and the reader can easily see how this applies to a
rank-n tensor.

Let bi be an arbitrary complete set of basis vectors in x-space. As shown in the notes following (6.2.8)
there exists a unique set of dual ("reciprocal") basis vectors bi (also in x-space) such that bi• bj = δij,
where we now use the Standard Notation bn equations of (7.18.6). Consider then the following expansion
of rank-3 tensor A with contravariant components Aabc,

A = Σijk αijk (bi⊗bj⊗bk) where (bi⊗bj⊗bk ...)abc = (bi)a (bj)b (bk)c (E.1.2)

Aabc = Σijk αijk (bi⊗bj⊗bk ...)abc = Σijk αijk (bi)a (bj)b (bk)c . (E.2.1)

As we did in Example 3 (2.9.5) for en, we declare the bn vectors to be "contravariant by definition" by
writing the rule (b'n)i ≡ Rij(bn)j. So then the vectors bn are true rank-1 tensors under x' = F(x).
According to Section 7.1, the outer product (bi)a (bj)b (bk)c is a true rank-3 contravariant tensor under
F. Then (E.2.1) says that Aabc is a linear combination of these tensors, which we know is a tensor
because the sum of two tensors of some type is a tensor of the same type.
The expansion coefficients αijk can be obtained by dotting both sides of (E.2.1) with (bi'⊗bj'⊗bk')
and using (E.1.5),

(bi'⊗bj'⊗bk') • (bi⊗bj⊗bk) = bi'• bi bj'• bj bk'• bk = δi'iδj'jδk'k . (E.2.2)

The result is then

A • (bi'⊗bj'⊗bk') = {Σijkαijk (bi⊗bj⊗bk)} • (bi'⊗bj'⊗bk') = Σijk αijk δi'iδj'jδk'k = αi'j'k'

or unpriming indices,

αijk = A • (bi⊗bj⊗bk) .

But since both A and (bi⊗bj⊗bk) are elements of the triple direct-product space spanned by the vectors
(bi⊗bj⊗bk), we use the dot product of (E.1.5) to claim that

A • (bi⊗bj⊗bk) = Aabc (bi⊗bj⊗bk)abc = Aabc (bi)a (bj)b (bk)c .

The coefficients αijk may then be written in all these ways :

αijk = A • (bi⊗bj⊗bk) = Aabc (bi⊗bj⊗bk)abc = Aabc (bi)a (bj)b (bk)c (E.2.3)

283
 Appendix E: Tensor Expansions

where Aabc are the contravariant components of tensor A in x-space, and (bi)a are the covariant
components of vector bi in x-space. As noted in Example 3 of the previous subsection, the object αijk
transforms as a scalar, since it is the dot product of two direct-product-space vectors. This fact is
especially obvious from the last expression in (E.2.3) where all tensor indices are contracted so the result
must be a scalar according to the neutralization rule (7.12.1). So αijk is a set of 33 =27 scalars.

Two special cases attract our attention.

The (ui)a are axis-aligned basis vectors in x-space, as shown in (7.13.9) or (7.18.3). For these basis
vectors, one has (ui)a = δia and (ui)a = δia . If one considers expansion (E.2.1) with bn = un, then

A = Σijk αijk (ui⊗uj⊗uk)

where (ui⊗uj⊗uk ...)abc = (ui)a (uj)b (uk)c = δia δjb δkc (E.2.4)

and the coefficients are found to be

αijk = A • (ui⊗uj⊗uk) = Aabc δia δjb δkc = Aijk , (E.2.5)

so the scalar coefficients αijk are exactly the x-space contravariant components of the tensor A. If this
seems paradoxical, see Example 1 above where we found that scalar s = r • ^x = x. Thus,

A = Σijk Aijk (ui⊗uj⊗uk) . (E.2.6)

On the other hand, if ei are the tangent base vectors in x-space (see Chapters 3), the dual vectors are the
ei and from (7.13.1), (ei)a = Sai = Ria and (ei)a = Sai = Ria . If one considers the expansion

A = Σijk αijk (ei⊗ej⊗ek)

where (ei⊗ej⊗ek ...)abc = (ei)a (ej)b (ek)c = Ria Rjb Rkc (E.2.7)

then the coefficients are found to be

αijk = A • (ei⊗ej⊗ek) = Aabc (ei)a(ej)b(ek)c = Aabc Ria Rjb Rkc

= Ria Rjb Rkc Aabc = A'ijk (E.2.8)

and thus the scalar coefficients αijk in this case are exactly the x'-space contravariant components of
tensor A, as shown for M in (7.10.9). Thus,

A = Σijk A'ijk (ei⊗ej⊗ek) . (E.2.9)

Expansions like the above are the generalizations (to tensors of any rank) of these vector expansions
stated in (7.13.12),

284
 Appendix E: Tensor Expansions

A = Σiαi bi αi = bi • A // arbitrary basis

A = ΣiAi ui // axis aligned unit vectors

A = ΣiA'i ei // tangent base vectors (E.2.10)

where we continue to write rank-1 tensors (vectors) in bold font: A.

To summarize, here is the general rank-n tensor expansion for an arbitrary basis, and then for the two
specific bases just discussed:

A = Σijk... αijk... (bi⊗bj⊗bk...) αijk... = Aabc... (bi)a (bj)b (bk)c...

A = Σijk... Aijk... (ui⊗uj⊗uk...) Aijk... = contravariant components of A in x-space

A = Σijk... A'ijk... (ei⊗ej⊗ek...) A'ijk... = contravariant components of A in x'-space

Tilting labels (E.2.11)

In the above discussion, suppose we make the swap bi → bi . We quickly trace the steps:

A = Σijk αijk (bi⊗bj⊗bk) where (bi⊗bj⊗bk ...)abc = (bi)a (bj)b (bk)c (E.2.1)

Aabc = Σijk αijk (bi⊗bj⊗bk ...)abc = Σijk αijk (bi)a (bj)b (bk)c = outer product sum (E.2.1)

αijk = A • (bi⊗bj⊗bk) = Aabc (bi⊗bj⊗bk)abc = Aabc (bi)a (bj)b (bk)c = scalar (E.2.3)

A = Σijk αijk (ui⊗uj⊗uk) // first special case (E.2.4)

αijk = A • (ui⊗uj⊗uk) = Aabc (ui)a (uj)b (uk)cc = Aabc δia δjb δkc = Aijk (E.2.5)

A = Σijk αijk (ei⊗ej⊗ek) // second special case (E.2.7)

αijk = A • (ei⊗ej⊗ek) = Aabc (ei)a(ej)b(ek)c = Aabc Ria Rjb Rkc = A'ijk (E.2.8)

Here αijk = Aabc (bi)a (bj)b (bk)c = Aabc (bi)a (bj)b (bk)c is a different set of scalars compared with
the original αijk = Aabc (bi)a (bj)b (bk)c because in general bi ≠ bi. The sequence above shows that
everything goes through as before, and so we may expand tensor A in either of these two ways,

285
 Appendix E: Tensor Expansions

A = Σijk αijk (bi⊗bj⊗bk)

A = Σijk αijk (bi⊗bj⊗bk) . (E.2.12)

The second expansion is really a mixed-basis expansion as discussed in Section E.10 below. The main
point of this exercise is to show that one may "reverse the tilt" of the i summation index and still have a
viable tensor expansion. This is not an example of the "contraction tilt-reversal rule" (7.11.1) because the
object αijk is not a rank-3 tensor, it is a set of scalars. Moreover, index i is not a tensor index, it is a label
on the basis vector bi. Nevertheless, in this case tilt-reversal is allowed and in fact just serves to define a
new set of coefficients αijk
It turns out that the indices on αijk can be lowered by the matrix w'nm = bn • bm which appears in
(7.18.6). To show this, consider

αijk = Aabc (bi)a (bj)b (bk)c = Aabc (bi)a (bj)b (bk)c // (E.2.3) then tensor tilt rule
so
w'siαijk = Aabc [w'sibi]a (bj)b (bk)c = Aabc [bs]a (bj)b (bk)c // bn = w'ni bi in (7.18.6)

= αsjk // as shown in (E.2.3) above (E.2.13)

which shows that w'** lowers an index on αijk. Despite this ability of w'** and w'** to lower and raise
indices on the object family αijk, each element of this family (such as αijk) is a scalar. When bn = en, we
find that g' then raises and lowers indices on αijk, but this does not make αijk a rank-3 tensor, because
αijk transforms as a scalar under x' = F(x). We apologize for constantly hammering on this point, but it
can be a source of confusion.
Obviously one can tilt any or all of the indices in expansion (E.2.1). In particular, tilting all indices
gives this alternate version of (E.2.11)

A = Σijk... αijk... (bi⊗bj⊗bk...) αijk... = Aabc... (bi)a (bj)b (bk)c...

A = Σijk... Aijk... (ui⊗uj⊗uk...) Aijk... = covariant components of A in x-space

A = Σijk... A'ijk... (ei⊗ej⊗ek...) A'ijk... = covariant components of A in x'-space

Orthonormal basis (E.2.14)

If the basis vectors bi happen to be orthonormal, as defined by bi• bj = δi,j then bi = bi because the
dual basis is unique. In this case w'ij = bi• bj = δij so the coefficient αijk is unchanged if any or all
indices are lowered, see (E.2.13).
An example of orthonormal basis vectors arises if bi = ^e i ≡ ei/|ei| = ei/h'i and ^
en • ^
e m = δn,m . In
this case we find from (7.18.1) that

g'nm = en • em = ^
en • ^
e m h'mh'n = δn,m h'mh'n = h'n2δn,m (E.2.15)

286
 Appendix E: Tensor Expansions

so x'-space has the diagonal metric tensor g'nm = h'n2δn,m . A specific example arises with polar
e =^
coordinates as shown in Fig (3.4.3) where ^ θ and ^
1 e = ^
r. 2
Since the dual basis is unique, one has bi = b = e = ^
i ^ i
e i †. In this case (E.2.1) and (E.2.3) state,

A = Σijk αijk (^
e i⊗^
e j⊗^
e k) αijk(any up/down) = Aabc (^
e i)a (^
e j)b (^
e k)c . (E.2.16)

On the other hand, one can expand A as in (E.2.9), using ei = h'i^

ei ,

A = Σijk A'ijk (ei⊗ej⊗ek) = Σijk [A'ijk h'i h'j h'k ](^

e i ⊗^
e j⊗^
e k) . (E.2.17)

Comparison of the last two equations shows that

αijk(any up/down) = Aabc (^

e i)a (^
e j)b (^
e k)c = A'ijk h'i h'j h'k // no implied sums (E.2.18)

Expansions on the unit versions of the tangent base vectors ^

e i are discussed more in Section E.8 below.

†
Is it really true that bi = bi ⇒ ^ ei = ^
e i where the latter two vectors are defined as unit vectors? This
will serve as a little check on our notation. Since g'ab = h'a2δa,b we know that g'ab = h'a-2δa,b. Then
g'nn = h'n2 and g'nn = h'n-2 . From (7.18.1) one has

|en| = g'nn = h'n en = g'ni ei = g'nn en = h'n-2 en

|en| = g'nn = (1/ h'n) en = h'n2 en
so
^
e n ≡ en/|en| = en h'n-1 = [h'n2en] h'n-1 = en h'n = en / |en| ≡ ^
en ⇒ ^
en = ^
en (E.2.19)

Thus one is free to move the ^

e n label up or down at will.

Tensor density expansions

If A is a tensor density of weight W, the general rule is to make this replacement:

A'ijk... → J WA'ijk... (E.2.20)

As justification for this rule, start with a regular tensor transformation for A,

A'ijk... = Rii' Rjj' Rkk'..... Ai'j'k'... (E.2.21)

The rule then gives

J W A'ijk... = Rii' Rjj' Rkk'..... Ai'j'k'... (E.2.22)

or
A'ijk... = J-W Rii' Rjj' Rkk'..... Ai'j'k'... (E.2.23)

287
 Appendix E: Tensor Expansions

which is the correct form for the transformation of a tensor density of weight W as in the example (D.1.4).

Using this rule, the expansion (E.2.9) of tensor density A would be written,

A = J W Σijk... A'ijk... (ei⊗ej⊗ek...) A'ijk... = contravariant components of A in x'-space

(E.2.24)
a result that is verified in (D.2.11). Taking components,

Aabc... = J W Σijk... A'ijk... (ei⊗ej⊗ek...)abc.. = J W Σijk... A'ijk... (ei)a (ej)b (ek)c ...

= J W Σijk... A'ijk... Ria Rjb Rkc... // see (7.18.1)

= J W Ria Rjb Rkc ...A'ijk (E.2.25)

which is just the inverse of (E.2.23). Applied to a vector A of weight W, the expansion (E.2.24) becomes

Aa = J WA'iei . (E.2.26)

E.3 Polyadic Notation

Some fields of study historically use "polyadic notation" as follows

(ABC...) ≡ A⊗B⊗C ... (E.3.1)

where the direct product notation was discussed above in Section E.1. It is sometimes a bit disturbing to
modern readers to see bolded vectors stacked directly against each other, but the direct product makes the
meaning clear. For arbitrary basis vectors, one would then have, for example,

(bibjbk...) ≡ bi⊗bj⊗bk ... . (E.3.2)

Sometimes this basis vector notation is compressed even more, to wit,

i j k ... ≡ (bibjbk...) ≡ bi⊗bj⊗bk ... (E.3.3)

although this notation seems to be mostly used when the bi are the unit vectors ui.
In all these notations, one must be aware that the symbols generally do not "commute". For example

i j = (bibj) = bi⊗bj ⇒ (i j )nm = (bibj)nm = (bi⊗bj)nm = (bi)n (bj)m (E.3.4)

nm nm nm n m nm
(j i ) = (bjbi) = (bj⊗bi) = (bj) (bi) ≠ (i j )

and therefore one cannot in general write i j = j i.

The general expansion (E.2.1) now appears as

288
 Appendix E: Tensor Expansions

A = Σijk... αijk... (bi⊗bj⊗bk...)

= Σijk... αijk... (bibjbk...)
= Σijk... αijk... (i j k...) (E.3.5)

where

αijk... = A • (bi⊗bj⊗bk...)
= A • (bibjbk...)
= A • (id jd kd ... )
= Aabc... (bi)a (bj)b (bk)c... (E.3.6)

where we have just made up a notation id to stand for the dual vector bi.
One can find further discussion of polyadic notation for example in Backus.

E.4 Dyadic Products

When two vectors A and B are combined in polyadic notation, the result is called a dyadic product (AB)
[ also known as a dyad or just a dyadic ]

(AB)ij ≡ AiBj // = (A⊗B)ij = component of the matrix A⊗B = AB . (E.4.1)

In this notation, the expansion (E.3.5) for a rank-2 tensor becomes

A = Σij αij (bibj) αij = Aab(bi)a (bj)b = Aab (bibj)ab . (E.4.2)

Notice from (7.1.1) that the dyadic product (AB) is a rank-2 tensor if we assume that the underlying Ai
and Bi are the x-space contravariant components of tensorial vectors A and B (which we normally
assume). As a reminder, x-space need not be Cartesian. In Section E.7 it will be shown that the matrix
(AB)ij can be associated with an operator (AB) in the un basis so (AB)ij = <ui |(AB)| uj >, but this
interpretation is not necessary for what follows.

If vector A is replaced by the gradient operator ∇, one gets

(∇B)ij ≡ ∇iBj = ∂iBj , // dyadic notation (E.4.3)

but it is more common to define the operator (∇B) using a reverse dyadic notation

(∇B)ij ≡ ∇jBi = ∂jBi = ∂Bi/∂xj // reverse dyadic notation (E.4.4)

since the index order i,j on the far right matches the order on the left.

289
 Appendix E: Tensor Expansions

E.5 Matrix notation for dyadics (Cartesian Space)

In Cartesian space g = 1 so for any vector a, we have ai = ai.

If we think of b as a column vector with components bj, then the corresponding row vector bT contains
those same elements bj, so we set (bT)j = bj. Therefore one can express the dyadic product in this more
down-to-earth manner,

(ab)ij ≡ aibj = ai (bT)j = (abT)ij (E.5.1)

or
ab = abT . (E.5.2)

Here one knows that ab is a "dyadic" because there is no other meaning for two bolded column vectors
abutting each other with no intervening operator, so no special notation like [ab] is needed to indicate that
ab is a dyadic. The object abT on the other hand has a well-defined meaning in matrix algebra,

abT =
⎛ a1 ⎞ (b b ) = ⎛ a1b1 a1b2 ⎞ = a matrix // same as matrix ab = a⊗b (E.5.3)
⎝ a2 ⎠ 1 2 ⎝ a2b1 a2b2 ⎠

and one sees that in fact

(ab)ij = (abT)ij = aibTj = aibj . (E.5.4)

Meanwhile, the object aTb is just a number,

⎛ b1 ⎞ = a b + a b = a • b .
aTb = (a1 a2) b // g = 1 . (E.5.5)
⎝ 2⎠ 1 1 2 2

This transpose notation can then be applied to the dyadic expansion of a 2x2 matrix A,

A = Σij αij bibj = Σij αij bibjT = α11 b1 b1T + α12 b1 b2T ... . (E.5.6)

In the special case that the bi are the unit vectors ui, and assuming N = 2 dimensions, one has

A = Σnm Anm unum = Σnm Anm unumT = A11 u1 u1T + A12 u1 u2T + A21 u2 u1T + A22 u2 u2T
= a matrix with A12 in the upper right corner (E.5.7)

where un is a column unit vector and unT is the corresponding row unit vector. For example,

1 0 1
u1u2T= ⎛⎝ 0 ⎞⎠ ( 0 1) = ⎛⎝ 0 0 ⎞⎠ . (E.5.8)

Obviously this matrix visualization is valid for any dimension N, not just N=2. For rank n > 2, however,
this transpose-of-vector concept does not conveniently generalize. For n=3 the object uaubuc would be a

290
 Appendix E: Tensor Expansions

cube of zeros with a single 1 located at coordinates a,b,c, and so on for n > 3. One cannot write this as
uaubucT for example.

In associating of M = abT with a rank-2 tensor, we must limit our interest to Cartesian space where g = 1.
That is because abT is the single matrix shown above in (E.5.5). If g ≠ 1, then there are four different
matrices Mij, Mij, Mij and Mij and the notation M = abT cannot support this fact. When g = 1 all four
matrices are the same matrix, and the x-space dot product is as shown in (E.5.5). g' ≠ 1 is still allowed.

Ambiguity of ab:

There is a certain ambiguity in (E.5.2) that ab = abT. Writing A = ab = abT, we imply that "ab" is the
name of a matrix, an alternate to the name A. This matrix A = ab has matrix elements Aij = (ab)ij =
aibj. On the other hand, in (E.3.1) we might say that B = ab = a⊗b which is a vector in a tensor product
space. So this version of the object ab is not a matrix, it is a tensor product space vector. So A and B are
different types of objects both indicated by the notation ab. However, the tensor components of the rank-2
tensor B are the same as the matrix elements of the matrix A. That is because (B)ij = [a⊗b]ij = aibj ,
being an outer product of two vectors. In Dirac notation described below in (E.7.4) we would write the
objects A and B in this manner which stresses the distinction,

A = |a><b| = the name of a matrix abT (or the corresponding operator in the Dirac Hilbert space)

B = |a> ⊗ |b> = a vector in a tensor product space. (E.5.9)

The matrix elements and tensor components (in the ui basis) would be

Aij = <ui| A | uj> = <ui| a><b| uj> = aibj

Bij = [<ui| ⊗ <uj| ] [ |a> ⊗ |b> ] = <ui|a><uj|b> = aibj . (E.5.10)

E.6 Large and small dots used with dyadics (Cartesian Space)

Sometimes a small-size dot • is used to indicate the action of a dyadic (matrix) on a vector. If A is a
dyadic (same symbol for matrix), and if c and d are vectors, then one defines:

A • c ≡ Ac = a column vector ⇒ (A • c)i = (Ac)i = Aijcj ⇒ A • c = ΣijAijcj ui

c • A ≡ cTA = a row vector ⇒ (c • A)i = (cTA)i = cjAji ⇒ c • A = ΣijcjAji ui
d • A • c = dTAc = a number = ΣijdiAijcj . (E.6.1)

It then follows that, for the particular dyadic A = ab ,

(ab) • c ≡ (ab) c = (abT)c = a(bTc) = a (b • c) = (b • c) a = a column vector

c • (ab) ≡ cT (ab) = cT(abT) = (cTa) bT = (c • a) bT = a row vector
d • (ab) • c = dT (ab) c = dTabT c = (dTa)( bT c) = (d • a)(b • c) = a number . (E.6.2)

291
 Appendix E: Tensor Expansions

Here is more detail on the first line of the above group showing a skeletal matrix structure,

(ab) c = (a bT)c = abTc = a(bTc) = a(b•c)

x x ⎛c1⎞ ⎛ a1 ⎞ ⎛c1⎞ ⎛ a1 ⎞ ⎛c1⎞ ⎛ a1 ⎞ ⎛c1⎞ ⎛ a1 ⎞
{ ⎛⎝ x x ⎞⎠ } c = { a (b1 b2)} c = a (b1 b2) c = a { (b1 b2) c } = a b•c
⎝ 2⎠ ⎝ 2⎠ ⎝ 2⎠ ⎝ 2 ⎠ ⎝ 2⎠ ⎝ 2 ⎠ ⎝ 2⎠ ⎝ 2 ⎠
(E.6.3)

The same small dot is used to indicate the product of two dyadics, which is to say, matrix multiplication

A•B ≡ AB . (E.6.4)

Regarding this small size dot • : (1) from a matrix algebra point of view, it is completely superfluous
except in the case c • A ≡ cTA ; (2) it is completely different from the dot • used in bTc = b•c. The
next Section provides an explanation of the small dot as part of an operator interpretation for dyadics.

E.7 Operators and Matrices for Rank-2 tensors: the bra-ket notation (Cartesian Space)

We continue with the assumption that x-space is Cartesian, g = 1. The discussion below applies only to
rank-2 tensors. One of our goals is to make contact with the dyadic discussion of Morse and Feshbach,
one of the few places dyadic notation appears in a general mathematical physics book.

Operator concept. As discussed in Section 5.10, x-space and x'-space of Picture A are both N-dimensional
real Hilbert Spaces, with a scalar product indicated by the large dot •, and one can regard V as a vector in
either space. The first line of (E.2.10) with A replaced by V states this expansion,

V = Σiαi bi αi = bi • V // arbitrary basis (E.2.10)

which we now rewrite, renaming the coefficients,

V = Σi[V(b)]i bi [V(b)]i = bi • V (E.7.1)

Moreover, one can regard a rank-2 tensor A as an "operator" in this Hilbert space, using notation (E.5.6),

A = Σij [A(b)]ij bibjT . (E.7.2)

In (E.2.10) and (E.5.6) the coefficients in these two expansions were called αi and αij, so here we are
providing more descriptive names for these coefficients.

Application of (bn)T on the left of (E.7.2) and bm on the right, and then a double use of (bn)Tbi = bn • bi
= δn,i (see 7.18.6) gives,

[A(b)]nm = (bn)T A bm . (E.7.3)

292
 Appendix E: Tensor Expansions

Here, one regards A as an operator in the x Hilbert space, whereas [A(b)]nm is a "matrix" which is
associated with the operator A in the particular bn basis. If bn = un, then [A(u)]nm = Anm which is a
particular matrix for this particular un basis. More generally,

[A(b)]nm = (bT)n A bm = [(bT)n]i Aij [bm]j = [bm]i Aij [bm]j

which is in general a completely different matrix formed by linearly combining elements Aij.

Bra-ket Notation. For the author of this document, the bra-ket notation commonly used in quantum
mechanics (Paul Dirac 1939) provides a useful way to look at a rank-2 tensor A as an operator. It is true
that in quantum mechanics one usually deals with infinite-dimensional Hilbert spaces and complex
numbers, but the formalism applies just as well to real Hilbert spaces with finite dimensions (used for
example to study "spin"). Here is how the bra-ket notation works:

bi → |bi> // vector
biT → <bi| // transpose vector

bi → |bi> // vector
(bi)T → <bi| // transpose vector

abT (E.5.3) → |a><b| // a matrix (outer product)

a Tb = a • b (E.5.5) → <a | b> // a number (inner product)

a • b = b • a (7.4.14) → <a | b> = <b | a> for any a,b // real Hilbert Space

V = Σi [V(b)]i bi (E.7.1) → |V> = Σi [V(b)]i |bi> // vector expansion ...

[V(b)]i = bi • V (E.7.1) → [V(b)]i = <bi|V> // and coefficients

bi • bj = δij = δi,j (7.18.3) → <bi|bj> = δi,j // orthogonality

i j i
= <bj|b > = <bi|b > = <bj|b >

Σi bi biT = 1 (7.18.6) † → Σi |bi><bi| = 1 = Σi |bi><bi| // completeness

A = Σij [A(b)]ij bibjT (E.7.2) → A = Σij [A(b)]ij | bi> <bj| // tensor expansion ...

[A(b)]ij = (bi)T A bj → [A(b)]ij = <bi | A | bj > // and coefficients

bi• (Abj) = bj • (AT bi) (7.9.17) → <bi | A | bj > = <bj | AT | bi > // covariant transpose

(E.7.4)

†
What appears in (7.18.6) for completeness is (bn)i(bn)j = δij, but since g = 1 we can write this with
indices down as (bn)i(bn)j = δi,j which in matrix notation says Σi bi biT = 1 .

293
 Appendix E: Tensor Expansions

In this bra-ket notation, the N |bi> are a set of basis vectors which span an N-dimensional real Hilbert
Space, while <bi| span the so-called adjoint (or transpose in our case) Hilbert Space. One then refers to
[A(b)]ij = <bi | A | bj > as "the matrix element of the operator A in the bi basis ", since the expansion
(E.7.2) was A = Σij [A(b)]ij bibj = Σij [A(b)]ij bi⊗bj which is an expansion in the bi basis. In
general, |bi> and |bi> are different vectors because bi and bi are different. For example en = g'nm em and
our restriction to g = 1 certainly allows g' ≠ 1. [ Do not confuse with (bi)a = (bi)a being the same; here a
is a tensor index raised and lowered by g = 1, whereas i is a basis vector label and these are raised and
lowered by w' as shown in (7.18.6).]

In this notation, based on what was presented earlier, one can write,

Anm = <un | A | um > = the x-space components of tensor A (basis un) raise/lower with g

A'nm = <en | A | em > = the x'-space components of tensor A (basis en) raise/lower with g'

[A(b)]nm = <bn | A | bm > = the matrix of A in the bn basis raise/lower with w

(E.7.5)

In the first of these three lines, one can raise and lower indices with gab and gab on both sides of the
equation, but we are taking g = 1 so up and down here do not matter. On the second line this can be done
with g'ab and g'ab. Eq (7.18.6) shows that bn = w'nm bm and conversely bn = w'nmbm where w'nm is the
metric tensor g'nm one would get for some underlying transformation Fb which causes bn to be its tangent
base vectors en. So, on the third line above we can raise and lower indices on each side with wab and wab
where w'nm ≡ bn • bm as in (7.18.6).
Notice in the three equations of (E.7.5) that the operator A between the vertical bars is the exact same
operator in each case. The matrices are different not because the operator has changed, but because the
basis vectors are different.
In a more consistent notation one might write Anm = [A(u)]nm and A'nm = [A(e)]nm .

Bases are related by a transformation. Consider again,

[A(b)]nm = <bn | A | bm > = (bn)T A bm = [bn]i Aij [bm]j = <bn|ui><ui|A|uj><uj|bm> . (E.7.6)

We lower index m on both sides (using w'ab as noted above) and reverse the j tilt to get

[A(b)]nm = <bn | A | bm > = (bn)T A bm = [bn]i Aij [bm]j = <bn|ui><ui|A|uj><uj|bm> . (E.7.7)

One could then define the following tensor-like object,

Bni ≡ [bn]i . (E.7.8)

The first index on B is raised and lowered by w', while the second is raised and lowered by g, so this
object is a bit like R and S in its non-tensor nature. Lowering n and raising i then gives

294
 Appendix E: Tensor Expansions

Bni = [bn]i = (BT)in , (E.7.9)

where we use the covariant transpose of a tilted matrix described in (7.9.3). One then has

[A(b)]nm = Bni Aij(BT)jm . (E.7.10)

Since all the matrices are tilted the same way and summed indices are contractions, this is one of the
"legal" Standard Notation matrix multiplication forms like (7.8.6) and we then write,

A(b) = BABT or more precisely [A(b) = BABT ]SN,dt (E.7.11)

where SN,dt means Standard Notation, down-tilt, as described below (7.8.6). The matrix equation A(b) =
BABT shows that the [A(b)]nm are related to the Aij by a "congruence transformation" with a matrix Bni
= [bn]i whose rows are the basis vectors bn. When bm = um , matrix B is the identity matrix, and when bm
= em one has Bni = [en]i = Rni, so that B = R in this case. In Standard Notation the general R matrix is
"covariant real orthogonal", [RRT = 1]SN,dt and [RT = R-1]SN,dt (see (7.9.3) and following text) so in
fact one has for the bm = em basis,

A(e) = B A BT = R A RT = R A R-1 = R A S . (E.7.12)

Specifically in this case,

[A(e)]nm = RniAijSjm = RniRmjAij = A'nm . (E.7.13)

which is the expected result looking at the second line of (7.5.8).

Comment: Recall that in Developmental Notation one has RRT = 1 and RT = R-1 only when R is a
rotation, whereas in Standard Notation one has RRT = 1 and RT = R-1 for any R matrix, see (7.9.3). We
used this fact above to show that a basis change from basis un to bn on operator A can be thought of as a
congruence transformation by a matrix B whose rows are the vectors bn. This is a standard concept in
linear algebra where one operates in Cartesian x-space.

More on bra-ket notation and its relation to the small dyadic dot.

Consider the following facts, where AT is the covariant transpose as in (7.9.3) and (7.9.17),

<d | A | c > = dT A c = dT [A c ] = <d |Ac >

<d | A | c > = dT A c = [ dT A] c = [AT d]T c = <ATd | c> (E.7.14)

where

|(Ac) > = a new Hilbert space vector which results when operator A is applied to |c> = A|c>

<(ATd) | = a new transpose Hilbert space vector which results when A is applied to <d| = <d|A .

295
 Appendix E: Tensor Expansions

(E.7.15)

Detail: In covariant notation [dTA] = [AT d]T because [dTA]i = ([AT d]T)i. To verify this,

[dTA]i = (dT)jAji = djAji = Ajidj

([AT d]T)i = [AT d]i = (AT)ijdj = Ajidj

So one has this general idea that

<d | A | c > = <d |Ac > = <ATd | c>

A | c > = |(Ac)> , <d | A = <(ATd) | . (E.7.16)

In this last line, the isolated A's are the same operator A sitting in the Hilbert space. This operator can
"act" either to the right or to the left as shown. The object |(Ac)> ≡ |e> is some different vector in the
Hilbert space (different from |c>), call it |e>, and the grouping (Ac) labels this vector. Similarly, <(ATd) |
is some vector <f| in the transpose Hilbert space. The distinction between A as an abstract operator in the
Hilbert space, and the A in (Ac) and (ATd) = (dTA)T as vectors in the Hilbert space is a subtle one. It is
just this distinction that is implied by the small dot in the dyadic notation discussed in the previous
section, and here is the correspondence between the dyadic notation and the bra-ket notation:

A • c = Ac d • A = (ATd)T = dTA d • A • c = dTAc A•Bc

A | c > = |Ac> <d | A = <(ATd)| <d | A | c > = <d | Ac > AB| c > .
(E.7.17)

In the rightmost column operator B is applied first to |c> to get vector |(Bc)>, and then operator A is
applied to |(Bc)> to give yet another vector | (ABc)>. In bra-ket notation the product of two abstract
operators is given just as AB, but in dyadic notation it is written A • B.

Dyadics as operators. According to the above discussion, one can regard a dyadic (AB), being a rank-2
tensor, as an operator and not as a matrix. The matrix Tnm = (AB)nm = AnBm is specific to the un basis in
x-space,

Tnm = (AB)nm = <un |(AB)| um > = (un)T A BT um

= [(un)T]a Aa (BT)b [um]b = δna Aa Bb δmb = AnBm . (E.7.18)

In the generic bn basis one has

[(AB)(b)]nm = <bn |(AB)| bm > . (E.7.19)

It is to emphasize this operator view of a dyadic that Morse and Feshbach use fancy letters like U to
represent dyadics. Then their small-dot notation U • B emphasizes the idea of an operator acting on a
vector, equivalent to U| B>. Here are a few samples from their Section 1.6 on dyadics. Under each clip we

296
 Appendix E: Tensor Expansions

have tried to relate the stated equation(s) to our notation above. Like us, Morse and Feshbach are working
in a Cartesian x-space, so up and down indices are the same for x-space objects. Note that un = un when g
= 1. The M&F symbols U and an are our A and un :

p 55
A • b = Σmn um Amn bn b•A = Σmn bm Amn un (E.6.1)
or A |b> = Σmn |um><um|A|un><un|b> <b|A = Σmn<b|um><um|A|un><un| (E.7.20)

p 55
A = Σnm unAnm um (E.5.7) (E.7.21)

A-1 • A = A • A-1 = 1 // A-1 defined if matrix [A(u)]mn = Amn is invertible (E.7.22)

p 57

0•f =0 1•f= f; 1 = Σn ununT (E.7.4) with bn = un

or
0 | f> = |0> ; 1 | f> = | f> ; 1 = Σn | un><un| (E.7.4) (E.7.23)

p 59
Us = Σnm [Us]nm unum = [Us]11u1u1 + [Us]12u1u2 + ... = [Us]11 i i + [Us]12i j + ..
⎛ Ax Bz By ⎞
where i = u1, j = u2 , k = u3 and where [Us]ij = ⎜ Bz Ay Bx ⎟ = a symmetric matrix (E.7.24)
⎜ ⎟
⎝ By Bx Az ⎠

Notice the impressive name "idemfactor" for the identity operator 1 = Σi | ai><ai| = Σi aiaiT = Σiaiai
where Σiaiai uses the dyadic notation (E.5.2) .

Comment: A favorite bra-ket notation trick is to obtain useful results by inserting 1 = Σn | un><un | in
opportune places, as in (E.7.20) above :

297
 Appendix E: Tensor Expansions

A |b> = 1 A 1 |b> = [Σm | um><um | ] A [Σn | un><un | ] |b> = ΣnΣm | um><um |A| un><un |b>
= ΣnΣm | um> Amn bn = Σnm Amn bn |um> (E.7.25)

The vector A |b> = |Ab> has thus been expanded on the basis | um>.

E.8 Expansions of tensors on unit tangent base vectors: M and N

We start with the general Picture A (and later specialize to orthogonal coordinates),

(E.8.1)

In (E.2.9) it was established that one can expand a tensor A on the tangent base vectors en as

A = Σijk... A' ijk... (ei⊗ej⊗ek...) A' ijk... = contravariant components of A in x'-space

A' ijk... = Rii'Rjj'Rkk'...... A i'j'k'... . // tensor transformation rule (E.8.2)

We know from (7.13.6) that en = |en| ^

e n = h'n ^
e n so the above expansion for tensor A can be written

A = Σijk... { h'ih'jh'k......A' ijk...} (^

e i⊗^
e j ⊗^
e k...)

^
≡ Σijk... [A(e)]ijk... (^
e i⊗^
e j ⊗^
e k...) , (E.8.3)

where the unit-vector expansion coefficients are given by

^
[A(e)]ijk... = { h'ih'jh'k......A' ijk...} // ≡ A' ijk... see (E.8.5)

= h'ih'jh'k...... Rii'Rjj'Rkk'...... A i'j'k'...

= (h'i Rii')( h'j Rjj')( h'k Rkk') ..... A i'j'k'... (E.8.4)

Note that we have not at this point assumed that the {^

e i} unit vectors are orthonormal, so g' can be non-
diagonal.

298
 Appendix E: Tensor Expansions

Coefficient notation

In a curvilinear coordinates application of these expansions, the expansion coefficients are usually written
in the following manner,

^
[A(e)]ijk... = A' ijk... = Ax'ix'jx'k ......
(E.8.5)

where the x'n are the names of the coordinates. For example, for a rank-4 tensor in spherical coordinates
with coordinates x'1 = r, x'2 = x'3 = θ and x'3 = φ one might write

^
[A(e)]2213 = A'2213 = Aθθrφ . (E.8.6)

The script notation is the same shorthand we used in (7.13.12) where V = V'n en = V'n ^ e n . Note that
n ijk...
there is a prime on V' and a prime on A' above. The reason for this prime is that these quantities
are associated with the curvilinear coordinates (variables) which in Picture A and B are called x'i . For
example we have V'n = h'n V'n . In the Moon & Spencer Picture of (14.11) with u-space (metric tensor g)
on the left, the curvilinear coordinates are called ui. Since they then have no prime, we would use
unprimed scripted variables. For example, Vn = hn Vn . This is the convention of Chapter 14, but in these
Appendices we are using Pictures A and B.

^
As shown below, one may interpret [A(e)]ijk... = A'ijk... as the contravariant components of tensor
A with respect to a certain transformation FM, and in this sense the up and down index position is
significant. However, when we later assume the ^
e i are orthonormal, that forces g'ij to be diagonal and
forces the condition ^
ei = ^
e i as shown below (E.2.14). Then the up or down index position on
^
[A(e)]ijk... makes no difference in expansion (E.8.3). In the spherical coordinates example where the
^
e i are orthonormal, we would normally write the above coefficient as Aθθrφ.

Matrices M and N

It is convenient now to define some modified R matrices as

Mab ≡ h'a Rab (E.8.7)

so that (E.8.4) becomes

^
[A(e)]ijk... = Mii'Mjj'Mkk'...... A i'j'k'... . (E.8.8)

As shown below, these matrices M are the R-matrices of a transformation called FH that takes x-space to a
new x"-space in which the tangent base vectors are the unit vectors ^
e n, but we defer that interpretation to
first get some facts laid out.

299
 Appendix E: Tensor Expansions

Defining Nab to be the inverse of Mab, one has (verified just below),

Nab ≡ h'b-1Sab = h'b-1Rba (E.8.9)

so the inversion of (E.8.8) is given by

^
Aijk... = Nii'Njj'Nkk'...... [A(e)]i'j'k'... . (E.8.10)

To verify that this N is the correct inverse of M, use (E.8.7) and (E.8.9) to show

MakNkc = (h'a Rak)( h'c-1Rck) = (h'a/ h'c) Rak Rck = (h'a/ h'c)δac = δac (E.8.11)

making use of orthogonality rule #3 of (7.6.4), Rak Rck = δac . M and N can be written in terms of the en
and en vectors as follows,

Mni ≡ h'n Rni = h'n(en)i // (7.13.1)

Nin = h'n-1Rni = h'n-1(en)i = (^

e n)i , // (7.13.1) (E.8.12)

which says that

Nin = { ^
e1 , ^
e 2, .... } . (E.8.13)

Thus the columns of Nin are the unit tangent base vectors.

Matrix H and x"-space

In the discussion above one has x-space with basis vectors un and x'-space with basis vectors en (the
tangent base vectors). It is useful then to define x"-space as the space whose basis vectors are the ^
e n unit
vectors, which are generally not orthogonal. The relation (E.8.7) Mab ≡ h'a Rab can be written as a down-
tilt matrix equation,

M = HR where Hij ≡ diag(h'1, h'2.....) . (E.8.14)

It then follow that

N = M-1 = R-1H-1 = SH-1 . (E.8.15)

Finally, note that

x = xiui = x'iei = x'i(h'i^

e i) = x"i^
ei ⇒ x"i = h'i x'i
so
x" = H x' . (E.8.16)

300
 Appendix E: Tensor Expansions

This shows that the transformation FH from x'-space to x"-space is linear with matrix H,

x" = FH(x') = H x' . (E.8.17)

Eq. (7.5.2) then says that the corresponding R-matrix is (RH)ik ≡ (∂x"i/∂x'k) = Hik so RH = H and then
SH = RH-1 = H-1.

We can now show all three spaces in the same Picture as follows,

(E.8.18)

This picture shows that the transformation directly from x-space to x"-space is

FM(x) = H F(x) . (E.8.19)

Here F(x) is a (generally) non-linear transformation assumed to connect x'-space to x-space. This is then
concatenated with linear transformation H to get non-linear transformation FM(x).
^
We can now write A(e) as A" and restate (E.8.3) and (E.8.8) above as

A = Σijk... A" ijk... (^

e i⊗^
e j ⊗^
e k...) // rank-n tensor expanded on ^
e i ⊗^
e j⊗^
e k...

A" ijk... = Mii'Mjj'Mkk'...... A i'j'k'... // rank-n tensor transformation

Aijk... = Nii'Njj'Nkk'...... A" i'j'k'... // inverse of the above

A" i = Mij Aj // rank-1 tensor

A" ij = Mii'Mjj'A i'j' // rank-2 tensor

^
where A" ijk... ≡ [A(e)]ijk... ≡ A' ijk... (E.8.20)

The word "tensor" suddenly has a new meaning in the above equations. The equations indicate objects
being tensors with respect to this non-linear transformation FM(x) whose linearized-at-a-point matrix is RM
= M = HR, where R is the linearized-at-a-point matrix version of F(x). H is the diagonal matrix of scale

301
 Appendix E: Tensor Expansions

factors h'i which are associated with g'ij in x'-space. The Aijk... are contravariant components of tensor
A in x-space, while A"ijk... are the corresponding contravariant components of A in x"-space, all with
respect to FM(x) and its matrix RM = M where, for example, dx" = M dx.
One can compare the last line of (E.8.20) to the first line of the generic (7.5.8) showing how a rank-2
tensor transforms. The second line of (7.5.8) then tells us the manner in which the mixed tensor Aab
transforms under FM :

A" ab = Maa' Mbb' Aa'b' = Maa'Aa'b'Mbb' . (E.8.21)

If we use the Standard Notation covariant transpose shown in (7.9.3) [RM = M], then Mbb' = (MT)b'b and
the above becomes

A" ab = Maa'Aa'b' (MT)b'b = [ MAMT ] ab

from which we conclude that

A" = MAMT // that is to say, [ A" = MAMT]SN,dt (E.8.22)

Comment:

This matrix result is a special case of the generic result (E.7.11) that A(b) = BABT when bn = ^ e n,
n n
but it is not obvious how this works out so here are some details. In order to compute B i ≡ [b ]i in
(E.7.8) one has to know bn. One sees from (7.18.6) that bn = w'ni bi where

w'nm = bn • bm = ^
en • ^
e m = (h'nh'm)-1 en • em = (h'nh'm)-1g'nm ⇒

w'nm = (h'nh'm)g'nm . // since wnm and wnm are inverses (E.8.23)

Therefore,

bn = w'nm bm = (h'nh'm)g'nm ^
e m = h'ng'nm em = h'n en ⇒

Bni ≡ [bn]i = h'n (en)i (E.8.24)

But (E.8.12) says Mni = h'n(en)i so we have shown that Bni = Mni so BABT = MAMT.

The metric tensor g" in x"-space appearing in Fig (E.8.18) is this,

g"ab = e"a• e"b = ^

ea • ^
e b = (h'ah'b)-1 ea • eb = (h'ah'b)-1 g'ab = w'ab , (E.8.25)

so g" = w' in the Comment above.

302
 Appendix E: Tensor Expansions

At this point the x and x' spaces are completely general, and none of R, RH = H, RM = M is a rotation
matrix. In the next Section, we shall specialize the above picture so that x-space is Cartesian with g = 1,
and x'-space has orthogonal curvilinear coordinates x' which means g' is a diagonal matrix. In this
scenario, F(x) is in general non-linear and so then is FM(x) = H F(x). Since the ^e i now form a frame of
orthonormal vectors, and since ui also form such a frame, one will not be surprised to find that M is now
a rotation which relates these two frame sets.

E.9 Application of Section E.8 to Orthogonal Curvilinear Coordinates

We now switch to Picture B (g=1) and assume that the x'i are orthogonal coordinates,

(E.9.1)

and our three-frame picture (E.8.18) above then becomes

(E.9.2)

In this situation, x-space is Cartesian with

gab = gab = δa,b g'ab = h'a2 δa,b g'ab = h'a-2 δa,b.

g"ab = (h'ah'b)-1 g'ab = (h'ah'b)-1h'a2 δa,b = δa,b // from (E.8.25) . (E.9.3)

Notice that g"ab = g"(x)ab = ^ e a(x) • ^

e b(x) = δab so x"-space is locally Cartesian at any point x. For
example, in polar coordinates one has ^r•^ θ = 0 at any location x. Since g = 1 and g" = 1, the up or down
index position on a tensor object is not significant.

Now, regardless of what R and S are, M is a "rotation", where we include in this term possible axis
reflections. To show this fact, first note from (E.8.7) and (7.5.9) that

303
 Appendix E: Tensor Expansions

Mjk = (h'j Rjk) = (h'j g'jaRak) = (h'j [δja h'j-2]Rak) = (h'j-1Rjk) = (h'j-1Rjk) . (E.9.4)

Using orthogonality rule #3 that Rab Rcb = δac = δa,c it then follows that,

MikMjk = (h'i Rik) (h'j-1Rjk) = h'ih'j-1 ( RikRjk) = h'ih'j-1 δi,j = δi,j . (E.9.5)

But MikMjk = δi,j is the Standard Notation statement that M is a "rotation", as was shown in (7.9.11).
In developmental notation this appears as [MMT = 1]DN, as shown there. Since M is a rotation, N must be
the inverse rotation since N = M-1. From (7.9.11) we then also know that

Mik = Mik (E.9.6)

which is the alternate form for Mik being a rotation, Again, M is the R-matrix RM = M for FM.

Relation between M and N. Looking at (E.9.5),

Mab Mcb = δa,c (E.9.5)

and knowing that

Mab (M-1)bc = δac = δa,c (E.9.7)

one concludes that (M-1)bc = Mcb . But (M-1)bc = Nbc so

Nbc = Mcb // note: this does not say that N = MT in standard notation, see (7.9.3) (E.9.8)

which can be verified from the above expressions for N and M. Since (E.9.6) must also be true for
rotation N, we find that

Mik = Mik = Nki = Nki // g = 1, g' = diagonal (E.9.9)

Interpretation of N and M. Since en = S un ( this is (3.2.4) and (3.2.1) with e'n = un) and since en = h'n^
en ,
it follows that

^
e n = h'n-1en = h'n-1 S un
or
(^
e n)a = h'n-1 Sab (un)b = h'n-1 Sab δnb = h'n-1 San = Nan (E.8.9) = Nabδbn = Nab(un)b
or
^
e n = N un and (^
e n)a = Nan . // ⇒ un = M ^
en (E.9.10)

Since the un are the Cartesian unit vectors, it seems intuitively obvious that the transformation that moves
this frame of orthonormal unit vectors {un} into the orthonormal frame {^ e n} must be a "rotation".

304
 Appendix E: Tensor Expansions

It was shown in (E.9.8) that Nan = Mna , therefore

Mna = Nan = (^
e n)a . (E.9.11)

The rotation matrix Nan = (^e n)a has the orthonormal basis vectors ^e n as its columns, while the rotation
matrix M a = ( e n) has the e n as its rows. Recall from (E.2.15) and (E.2.19) that orthonormal ^
n ^ a ^ e n implies
both ^
en = ^
e n and g'nm = h'n2δn,m = diagonal.

The relation un = M ^ e n can be written un = M(x) ^ e n(x) to emphasize that the rotation M(x) = RM(x) is
really a different rotation at every point x, since the ^
e n(x) vary with x. This is very different from a global
rotation which is the same at all points. For a global rotation R, F = R = linear and ∂iuj is a tensor. For RM
being a rotation which varies from point to point, FM is non-linear just as F defining the curvilinear
coordinates is non-linear, and ∂iuj fails to be a tensor under either F or FM.

Matrix form of rank-2 tensor transformation:

One implication of the above picture relates to tensor equations being covariant, as discussed in Section
7.15. If one has a tensor field equation in x-space,

Qadc(x) = Hab(x)Tbc(x) Bd(x) , (E.9.12a)

in which all the objects transform as tensors with respect to the underlying x" = FM(x) (and its linear
approximation M(x) as in dx" = M(x) dx), then the equation is covariant and takes the same form in x"-
space,

Q"adc(x") = H"ab(x")T"bc(x") B"d(x") . x" = FM(x) (E.9.12b)

Using our script notation explained in (E.8.5) and (E.8.20), we rewrite the x"-space equation above as the
third line below (primed script tensor names). The fourth line is then appropriate to the Moon and Spencer
Picture (14.1.1) where the curvilinear coordinates ui have no primes :

Qadc = HabTbc Bd x-space (general), Picture A or M&S (1.11)

Q'adc = H'abT'bc B'd x'-space (general), Picture A or B (1.11)
Q'adc = H'abT'bc B'd x"-space (orthogonal) , Picture A or B (1.11)
Qadc = HabTbc Bd x"-space (orthogonal) , Picture M&S (14.1.1) (E.9.13)

Comment: An x-space observer has axes un (Frame S) while an x"-space observer has axes ^ e n(x) (Frame
S"), and these two sets of observation axes are related by un = M(x) ^
e n(x) where M(x) is a rotation. If the
first equation describes something at location x in the realm of Newtonian mechanics, we expect the
equation to have the same form in both Frame S and Frame S" which are related by this local rotation
M(x). In other words, rotations are an invariance of Newtonian mechanics, and this means equations are
covariant with respect to rotations. The above example, which might apply to fluid dynamics, has this

305
 Appendix E: Tensor Expansions

covariance at each point x in the fluid, and it may happen that the rotation is a different rotation at
different points x, but it is always a rotation.

A simple interpretation of (E.9.13) is that the mutually orthogonal unit base vectors ^
e n form a rotated
Cartesian frame of reference, so the x-space equations must have the same form in this rotated frame.
This is why the first and last lines of (E.9.13) have the same form.

Example: Polar Coordinates. In polar coordinates now with ordering r,θ = 1,2 one has

S11 = (∂x/∂r) = cosθ x = rcosθ

S12 = (∂x/∂θ) = -rsinθ y = rsinθ
S21 = (∂y/∂r) = sinθ
S22 = (∂y/∂θ) = rcosθ (E.9.14)

cosθ -rsinθ cosθ sinθ

Sij = ⎛⎝ sinθ rcosθ ⎞⎠ Rij = ⎛⎝ -sinθ/r cosθ/r⎞⎠ R = S-1

-2
cosθ sinθ cosθ -sinθ/r 1 0
[g' = RRT]DN = ⎛⎝ -sinθ/r cosθ/r⎞⎠ ⎛⎝ sinθ cosθ/r ⎞⎠ = ⎛ 0 1/r2 ⎞ → g'ab =
⎛ h'r 0 ⎞
⎝ ⎠ ⎝ 0 h'θ-2 ⎠
so hr = 1 and hθ = r .

The N and M matrices may be computed as follows: [ recall M = HR and N = SH-1 from (E.8.14,15) ]

1 0 cosθ sinθ cosθ sinθ

Mab ≡ h'a Rab = ⎛⎝ 0 r ⎞⎠ ⎛⎝ -sinθ/r cosθ/r⎞⎠ = ⎛⎝ -sinθ cosθ⎞⎠ = Rz(-θ) (E.9.15)

cosθ -rsinθ 1 0 cosθ -sinθ

Nab ≡ Sab h'b-1 = ⎛⎝ sinθ rcosθ ⎞⎠ ⎛⎝ 0 1/r ⎞⎠ = ⎛⎝ sinθ cosθ ⎞⎠ = Rz(θ) (E.9.16)

^
Therefore, the relation between a rank-2 tensor's ^ e n-expanded-form components [A(e)]ij and the
Cartesian form components Aij is given (for polar coordinates) by (E.8.22) which says

^
A" = A(e) = A' = M A MT = M A MT (E.9.17)
or
⎛ Arr Arθ ⎞ = ⎛ cosθ sinθ ⎞ ⎛ A11 A12 ⎞ ⎛cosθ -sinθ⎞ // verified in Lai p 316 Problem 5.71
⎝ Aθr Aθθ ⎠ ⎝ -sinθ cosθ⎠ ⎝ A21 A22 ⎠ ⎝ sinθ cosθ ⎠

where recall that up or down index position has no significance in either x-space or x"-space and MT =
MT as in (7.9.5). That is to say, one has A12= A12 as well as Arθ = Arθ, so one can think of the above as
a down-tilt matrix equation.

306
 Appendix E: Tensor Expansions

Applications in Continuum Mechanics

1. In isotropic elastic stress analysis for states of plane stress and plane strain (see Lai p 251), the 3D
Cartesian stress tensor Tij has a simple form in which the upper left four components can be represented
as derivatives of a potential-like function called an Airy function φ, so that T11 = ∂22φ, T22 = ∂12φ, and
T12 = T21 = – ∂1∂2φ. In this case, equation (E.9.17) becomes

2
⎛ Trr Trθ ⎞ = ⎛ cosθ sinθ ⎞ ⎛ ∂2 φ – ∂12∂2φ ⎞ ⎛cosθ -sinθ⎞ // B = MAMT (E.9.18)
⎝ Tθr Tθθ ⎠ ⎝ -sinθ cosθ⎠ ⎝ – ∂1∂2φ ∂1 φ ⎠ ⎝ sinθ cosθ ⎠
where
∂1 = ∂/∂x1 = cosθ ∂r - (sinθ/r)∂θ
∂2 = ∂/∂x2 = sinθ ∂r + (cosθ/r)∂θ .

Here is a Maple computation of (E.9.18):

307
 Appendix E: Tensor Expansions

// verified in Lai p 264 (5.27.3) (E.9.19)

This then is a real-world example of using a rank-2 tensor in curvilinear coordinates expanded on the unit
tangent base vectors. The mentioned plane of strain or stress has Cartesian coordinates x1,x2 which are
converted to polar coordinates r,θ. The third Cartesian coordinate x3 is more or less ignored.

2. The relation between the stress tensor Tij and the infinitesimal strain tensor Eij for an isotropic
material is stated in Cartesian coordinate x-space as ( a form of Hooke's Law generalizing F = -kx),

Tij = λ tr(E)gij + 2μEij or the same thing: Tij = λ tr(E)gij + 2μEij

where tr(E) = Σi Eii = Eii with implied sum on i . (E.9.20)

Here λ and μ are called Lamé's constants and gij = gij = δi,j (see Lai p 208 (5.3.8)). Some components
are (in the case that E33 = 0),

T11 = λ(E11+ E22) + 2μE11 // 1,2,3 = x,y,z

T22 = λ(E11+ E22) + 2μE22
T12 = 2μE12 . (E.9.21)

With respect to our transformation FM, (E.9.20) is a "true tensor equation" (tr(E) = Ekk = Ekk is scalar
under rotations), so according to Section 7.15 it is "covariant" and in x"-space may be written

308
 Appendix E: Tensor Expansions

T"ij = λ tr(E")g"ij + 2μE"ij

or
T"ij == λ tr(E")δi,j+ 2μE"ij // g"ij = δi,j as in (E.9.3)
or
^ ^ ^
[T(e)]ij = λ [T(e)]kk δi,j + 2μ[E(e)]ij .
or
T'ij = λ tr(E')δi,j + 2μE'j tr(E') = Σi E'ii . // E = script E (E.9.22)

This notion of equation covariance was shown above in example (E.9.13). Recall the notation example
from (E.8.6) that, for spherical coordinates 1,2,3 = r,θ,φ,

^
[A(e)]2213 = A'2213 = Aθθrφ = Aθθrφ . (E.8.6)

Correspondingly, we would set T'11 = Trr and so on to write the components of (E.9.21) this way in
cylindrical coordinates, where 1,2,3 = r,θ,z,

Trr = λ [ Err+ Eθθ] + 2μ Err // Ezz = 0

Tθθ = λ [ Err+ Eθθ]) + 2μEθθ
Trθ = 2μ Erθ . (E.9.23)

This has the same form as (E.9.21) with 1→r and 2→θ. These equations are consistent with Lai p 264
(5.27.7) with λ = -EYv/[(v+1)(2v-1)] and 2μ = EY/(1+v). Statements of (E.9.23) are rare on the web, but
here is one instance (Victor Saouma draft Lecture Notes on Continuum Mechanics, 1998),

(E.9.24)

3. By way of contrast, the Cartesian-coordinates equation Eij = (∂iuj + ∂jui)/2, which relates strain
tensor Eij to the vector displacement u of a continuum particle, is not a "true tensor equation", so
Erθ ≠ (∂ruθ + ∂θur)/2. In fact, this relation is E = [(∇u)T + (∇u)] / 2 and (∇u) for polar coordinates is
computed in (G.6.6) and one ends up with Erθ = (∂ruθ + (1/r) ∂θur - uθ/r) / 2 .

E.10 Tensor expansions in a mixed basis

Recall the expansion (E.2.1) and (E.2.3) written for a general rank-n tensor A,

A = Σijk... αijk... (bi⊗bj⊗bk...) αijk... = A • (bi⊗bj⊗bk...) (E.10.1)

where αijk... are the coefficients of the expansion of A on the direct product basis shown. To make
explicit the fact that the coefficients depend on the choice of basis, one might write (one b for each index,
number of b's is the rank of the tensor) ,

309
 Appendix E: Tensor Expansions

αijk... = [ A(b,b,b...)]ijk... . (E.10.2)

The fact that the indices ijk... are "up" indicates that the b label stands for the bi basis and not bi. So here
is an example showing the generic expansion of a rank-3 tensor,

A = Σijk [ A(b,b,b)]ijk (bi⊗bj⊗bk) [ A(b,b,b)]ijk = A • (bi⊗bj⊗bk) . (E.10.3)

Earlier we used the simpler notation [A(b)]ijk for the above coefficient, but now we want to show all the
basis elements because now we want to consider a "mixed basis expansion" such as

A = Σijk [ A(b,e,u)]ijk (bi⊗ej⊗uk) [ A(b,e,u)]ijk = A • (bi⊗ej⊗uk) . (E.10.4)

This is a completely viable expansion since the b, e and u basis vectors are each a complete set within
their part of the direct-product space. To verify the validity of this expansion, consider :

[ A(b,e,u)]ijk = {A} • (bi⊗ej⊗uk)

= { Σi'j'k' [ A(b,e,u)]i'j'k' (bi'⊗ej'⊗uk')}• (bi⊗ej⊗uk)

= { Σi'j'k' [ A(b,e,u)]i'j'k' (bi'• bi) (ej'• ej) (uk'• uk)

= { Σi'j'k' [ A(b,e,u)]i'j'k' δi'iδj'jδk'k

= [ A(b,e,u)]ijk . (E.10.5)

In the case of a rank-2 tensor, one has the option of using the other notations discussed above,

A = Σij [ A(b,u)]ij (bi⊗uj) = Σij [ A(b,u)]ij (biuj)

direct product dyadic

= Σij [ A(b,u)]ij (biujT) = Σij [ A(b,u)]ij |bi><uj| (E.10.6)

matrix bra-ket
where

[ A(b,u)]ij = A • (bi⊗uj) = Aab (bi)a (uj)b = (bi)T A (uj) = < bi| A | uj > . (E.10.7)

One can of course use unit versions of the en basis vectors, ^

e n, and then one might write for example

^ ^
A = Σijk [ A(e,e,u)]ijk (^
e i ⊗^
e j⊗u k) (E.10.8)

where the hats are replicated into the superscript tensor label.

310
 Appendix F: Affine Connection

Appendix F: The Affine Connection Γcab and Covariant Derivatives

F.1 Definition and Interpretation of Γ : Γcab = qc • (∂aqb) = Rci(∂aRbi)

In this Section we shall use a modified Picture C in which the quasi-Cartesian space on the right is called
ξ-space instead of x(0)-space as in Picture C. The notation ξi for the coordinates of ξ-space seems
traditional in general relativity work where the Γ object appears frequently.

(F.1.1)

Recall from the discussion near (1.10) that the metric tensor G is a diagonal matrix whose elements are
independently +1 or -1. Since the metric tensor transforms as a rank-2 tensor, we know from the last line
of (7.5.8) (adapted from Picture A to Picture C1) that gab = RaiRbjGij. Since G is diagonal, we write
this as

gab = RaiRbiGii and gab = RaiRbiGii (F.1.2)

with a single implied sum on i. If G = 1, then the xi coordinates of x-space are "the curvilinear
coordinates" and the ξi are "the Cartesian coordinates".

In Picture C1 the tangent base vectors exist in ξ-space and we will call them qn. From (7.18.1) (adapted
from Picture A to Picture C1) we know both that (qn)i = Rni and that the dot product qn • qm = gnm
where g is the metric tensor in x-space. In general qn = qn(ξ). However, since x = F(ξ), we are free
instead to regard qn = qn(x), and that is what we shall do for Picture C1.

If one moves a small amount dx in x-space, qn(x) (a vector in ξ-space) will change by some small
amount. We are used to this idea in the Picture A context of Fig (3.4.3) where the x-space tangent base
vectors e = r ^
θ θ and e = ^
r both change if one takes θ→θ+dθ and r→r + dr in x'-space.
r

So, for a small change dxj in x-space, the change in qn(x) is given by,

d(qn)i = ∂j(qn)i dxj // ∂j ≡ ∂/∂xj

or
i i j ∂(fi) ∂fi
(dqn) = (∂jqn) dx . // ∂x(f ) = dx = dx = (∂xfi)
i
(F.1.3)

Since qn and (∂jqn) are both vectors in ξ-space, and since the qk are known to form a complete basis in ξ-
space, it must be possible to expand (∂jqn) on the qk with some appropriate coefficients, call them Γkjn :

311
 Appendix F: Affine Connection

(∂jqn) = Γkjn qk ⇒ (dqn)i = (∂jqn)dxa = Γkjn (qk)i dxa . (F.1.4)

The coefficients Γkjn are known as the affine connection. They measure how the tangent basis vectors
change in ξ-space as a function of x in x-space. One regards Γkjn(x) as an object associated with x-space,
though it is not a tensor object as (F.6.3) below shows.

Dotting the left equation into qm , using qm • qk = δmk as in (7.18.1), and then doing m→ k gives

Γkjn = qk • (∂jqn) = (qk)i(∂jqn)i = Rki(∂jRni) , (F.1.5)

where from (7.18.1) and (7.5.16) (adjusted from Picture A to Picture C1) we have used

∂xk ∂ξi
(qk)i = Rki = ∂ξi and (qn)i = Rni = ∂xn . (F.1.6)

Inserting the partial derivatives from (F.1.6) into (F.1.5) gives

∂xk ∂ξi ∂xk ∂2ξi

Γkjn = Rki(∂jRni) = ∂ξi ( ∂j ∂xn ) = ∂ξi ∂xj∂xn (F.1.7)

Notice that

∂2ξi ∂2ξi
(∂jRni) = (∂nRji) // since ∂xj∂xn = ∂xn∂xj (F.1.8)

Form (F.1.7) shows that:

Fact: Γkjn is symmetric on the lower two indices, so Γkjn = Γknj (F.1.9)

If we take k→λ, i→α, j→μ and n→ν, then (F.1.9) becomes

or
∂xλ ∂2ξα
Γλμν = ∂ξα ∂xμ∂xν

and this equation appears as Weinberg p 100 (4.5.1), with traditional Greek indices used for relativity.

We now restate some of the results above with more commonly used indices :

(∂aqn) = Γkan qk → (∂aqb) = Γcab qc

Γkjn = qk • (∂jqn) → Γcab = qc • (∂aqb)

312
 Appendix F: Affine Connection

so that

(∂aqb) = Γcab qc (F.1.4)

Γcab = qc • (∂aqb) (F.1.5)
Γcab = Γcba (F.1.9)

Γcab = Rci(∂aRbi) (F.1.5)

Γcab = – Rbi (∂aRci) (F.1.10)

The last line will be derived in the next Section.

One can convert the Γ-related equations from Picture C1 to Picture A,

For example, since the tangent base vectors in x-space of Picture A are called en,

(∂aqb) = Γcab qc → (∂'aeb) = Γ'cab ec

Γcab = qc • (∂aqb) → Γ'cab = ec • (∂'aeb) (F.1.11)

where Γ' = Γ'(x') and en = en(x').

F.2 Identities of the form (∂aRdn) = – Ren Rdm (∂aRem)

The identities (to be proven below) are :

Rdm (∂aRem) = – Rem (∂aRdm) 1

(∂aRdn) = – Ren Rdm (∂aRem) 2 ∂a ≡ ∂/∂xa

(∂aRdn) = – Ren Rdm (∂aRem) 3 . (F.2.1)

Corollary: The first identity above allows the alternate form for the affine connection in terms of R, as
quoted above in (F.1.10). Here we take d→c, m→i and e→b in identity #1,

Rdm (∂aRem) = – Rem (∂aRdm) // identity 1

Rci (∂aRbi) = – Rbi (∂aRci) // same with new indices

which then verifies (F.1.10).

313
 Appendix F: Affine Connection

Our context is:

. (F.2.2)

Proofs of identities 1,2,3: These identities are a simple consequence of the fact that RS = 1 which in
standard notation is written δcb = RcαRbα (orthogonality rule #3 in (7.6.4) ). So,

0 = ∂a(δde) = ∂a(RdmRem) = Rdm (∂aRem) + Rem (∂aRdm) . QED 1 (F.2.3)

Apply Σe Ren to both ends of (F.2.3) to get

0 = Ren Rdm (∂aRem) + (Ren Rem) (∂aRdm) = Ren Rdm (∂aRem) + δnm (∂aRdm)
= Ren Rdm (∂aRem) + (∂aRdn)

⇒ (∂aRdn) = – Ren Rdm (∂aRem) . QED 2 (F.2.4)

Alternatively, apply Σd Rdn to both ends of (F.2.3) to get

0 = (Rdn Rdm) (∂aRem) + Rdn Rem (∂aRdm) = δnm (∂aRem) + Rdn Rem (∂aRdm)
= (∂aRen) + Rdn Rem (∂aRdm)

⇒ (∂aRen) = – Rdn Rem (∂aRdm) now swap d and e:

⇒ (∂aRdn) = – Ren Rdm (∂aRem) . QED 3 (F.2.5)

F.3 Identities of the form (∂cgab) = – [gan Γ bcn + gbn Γacn]

The derivatives of the metric tensor are given by ( ∂c = ∂/∂xc)

(∂cgab) = – [gan Γ bcn + gbn Γacn] 1 (F.3.1)

(∂cgab) = + [gan Γncb + gbn Γnca] 2 (F.3.2)

(F.3.3)

314
 Appendix F: Affine Connection

Proof of 1: (∂cgab) = – [gan Γ bcn + gbn Γacn] (F.3.1)

Using (F.1.2) that gab = RaiRbiGii, the LHS of (F.3.1) becomes,

LHS = (∂cgab) = ∂c(RaiRbi)Gii = Rai(∂cRbi)Gii + Rbi(∂cRai)Gii . (F.3.4)

For the RHS, the Γ objects can be replaced by their alternate definitions from (F.1.10),

Γcab = – Rbi (∂aRci) // from (F.1.10)

Γbcn = – Rni (∂cRbi) // b→n then c→b then a→c
Γacn = – Rni (∂cRai) . // b→a (F.3.5)

The RHS of the claimed identity (F.3.1) may then be written

RHS = – gan Γ bcn – gbn Γacn

= {RakRnkGkk }{Rni (∂cRbi)} + {RbkRnkGkk }{Rni(∂cRai)}

= Rak(Rnk Rni) (∂cRbi) Gkk + Rbk(Rnk Rni) (∂cRai) Gkk

= Rakδki (∂cRbi) Gkk + Rbkδki (∂cRai) Gkk // orthog rule #2 of (7.6.4)

= Rai (∂cRbi) Gii + Rbi (∂cRai) Gii

= (∂cgab) // using (F.3.4) QED 1 (F.3.6)

Proof of 2: (∂cgab) = + [gan Γncb + gbn Γnca] (F.3.2)

Using (F.1.2) that gab = RaiRbiGii, the LHS of (F.3.2) becomes,

LHS = (∂cgab) = ∂c(RaiRbi)Gii = Rai (∂cRbi)Gii + Rbi (∂cRai)Gii . (F.3.7)

For the RHS, the Γ objects can be replaced by their primary definitions in (F.1.10),

Γcab = Rci(∂aRbi)
Γncb = Rni(∂cRbi) // c→n then a→c then i→k
Γnca = Rni(∂cRai) . // b→a (F.3.8)

The RHS of the claimed identity (F.3.2) may then be written

RHS = gan Γncb + gbn Γnca

= {RakRnkGkk}{Rni(∂cRbi)} +{RbkRnkGkk}{Rni(∂cRai)}

315
 Appendix F: Affine Connection

= Rak (Rnk Rni)(∂cRbi)Gkk + Rbk(Rnk Rni)(∂cRai)Gkk

= Rak δki(∂cRbi)Gkk + Rbk∂ki(∂cRai)Gkk // orthog rule #1 of (7.6.4)

= Rai (∂cRbi)Gii + Rbi (∂cRai)Gii

= (∂cgab) // using (F.3.7) QED 2 (F.3.9)

F.4 Identity: Γdab = (1/2) gdc [ ∂agbc + ∂bgca – ∂cgab]

The identity states that Γdab may be expressed entirely in terms of the metric tensor,

Γdab = (1/2) gdc [ ∂agbc + ∂bgca – ∂cgab] . (F.4.1)

Recall in our definition above, Γdab = Rdk(∂aRbk), that Γ was given in terms of R matrices.

The following corollary concerns summing of the upper Γ index with a lower one,

Γaan = (1/2) gad ∂ngad = (1/2)(1/g)∂ng = (1/ |g| ) ∂n( |g| ) . (F.4.2)

We shall prove the identity (F.4.1) first, and the corollary (F.4.2) after that.

Proof of Identity: Γdab = (1/2) gdc [ ∂agbc + ∂bgca – ∂cgab] (F.4.3)

We first rewrite (F.1.2) with new index names,

gab = RaeRbe Gee // sum on e

gdc = RdiRci Gii . // sum on i (F.4.4)

The first line below is computed from the first line in the pair above, then the next two lines below are
obtained by doing forward cyclic index permutations of the first line :

∂cgab = [Rbe (∂cRae) + Rae(∂cRbe) ]Gee

∂agbc = [Rce (∂aRbe) + Rbe(∂aRce) ]Gee
∂bgca = [Rae (∂bRce) + Rce(∂bRae) ]Gee . (F.4.5)

The last four lines can be appropriately inserted into the RHS of (F.4.3) to obtain

316
 Appendix F: Affine Connection

(RHS)dab = (1/2) gdc [ ∂agbc + ∂bgca – ∂cgab]

= (1/2) (RdiRci Gii) Gee *

[Rce (∂aRbe) + Rbe(∂aRce) + Rae (∂bRce) + Rce(∂bRae) – Rbe (∂cRae) – Rae(∂cRbe) ] . (F.4.6)
1 2 3 4 5 6

Due to the symmetry (∂iRje) = (∂jRie) noted in (F.1.8), terms 2 and 5 cancel as do terms 3 and 6, while
terms 1 and 4 are equal. Therefore,

(RHS)dab = (1/2) RdiRci Gii Gee * 2 Rce (∂aRbe)

= Rdi (Rce Rci) Gii Gee (∂aRbe)

= Rdi δei Gii Gee (∂aRbe) // orthog rule #1 of (7.6.4)

= Rde Gee Gee (∂aRbe)

= Rde(∂aRbe) // Gee = Gee = ±1

= Γdab // from (F.1.10)

= LHS of (F.4.3) QED (F.4.7)

Proof of Corollary: The corollary is this (g ≡ det(gab) )

Γaan = (1/2) gad( ∂agnd + ∂ngad – ∂dgan ) = (1/2) gad ∂ngad = (1/2)(1/g)∂ng = (1/ |g| ) ∂n( |g| ) .
1st 2nd 3rd 4th 5th
(F.4.8)
The 2nd expression is of course just (F.4.1) with c = b = n and implied sum on n.
The first and third terms of the 2nd expression cancel due to symmetry of g:

first term = gad ∂agnd = gda ∂dgan = gad ∂dgan = - third term (F.4.9)
a↔d gda = gad
so we then have just

Γaan = (1/2) gad(∂ngad)

which is the 3rd expression shown in (F.4.8).

To get the 4th expression, we must show that
gab(∂ngab) = (1/g)∂ng. Here are the steps to prove this fact, where g ≡ det(gab) :

317
 Appendix F: Affine Connection

(1) gab = (g-1)ab = cof(gab)T/det(gab) = cof(gab)/g ⇒ cof(gab) = g gab . (F.4.10)

(2) Apply gba and sum on b to get: gbacof(gab) = g gbagab = g δaa = g . (F.4.11)

(3) g = det(gab) = gabcof(gab) ⇒ ∂g/∂gab = cof(gab) = g gab * (F.4.12)

(4) ∂ng = ∂g/∂xn = (∂g/∂gab)( ∂gab/∂xn) = g gab (∂ngab) ⇒ (1/g) ∂ng = gab(∂ngab) (F.4.13)

* In Step 3, notice that ∂ cof(gab)/∂gab = 0 because cof(gab) is not a function of variable gab. The element
gab is one of those "crossed out" in obtaining the cofactor of gab. Despite this fact, cof(gab) is still an
object having indices a,b. In contrast, g = det(gab) = det(g**) has no indices!

It remains only to obtain the 5th expression in (F.4.8), and that is easy to show:

g-1/2 ∂n(g1/2) = g-1/2 (1/2) g-1/2 ∂n(g) = (1/2) (1/g) (∂ng) . g>0

[-g]-1/2∂n([-g]1/2) = [-g]-1/2 (1/2)[-g]-1/2∂n([-g]) = (1/2)[-g]-1 ∂n([-g]) g<0

which we summarize as

|g|-1/2 ∂n(|g|1/2) = (1/2) (1/g) (∂ng) (F.4.14)

We close this Section with the following observations:

Theorem: If x-space of Picture C1 of (F.1.1) has a metric tensor gij(x) whose elements are constants
independent of x, then Γcab ≡ 0. (F.4.15)

Corollary: If the x-space in Figure C1 of (F.1.1) is Cartesian (g =1) or quasi-Cartesian (g = G),

then Γcab ≡ 0. (F.4.16)

Proof: The identity (F.4.1) says Γdab = (1/2) gdc [ ∂agbc + ∂bgca – ∂cgab],
so if gij are constants, one must have Γdab = 0.

We shall often quote this Corollary with the phrase "Γ = 0 for a Cartesian space".

318
 Appendix F: Affine Connection

F.5 Picture D1 Context

Our main context of interest is Picture A,

. (F.5.1)

For the proof to be presented in Section F.6, it is useful to think of Picture A as the top part of this Picture
D1,

(F.5.2)

The relationship between the R's and S's are these

R = R' R-1 = R' S ⇒ R' = RR

S = R R'-1 = R S' . (F.5.3)

The tangent and reciprocal base vectors in ξ-space associated with transformations F and F ' are these,
based on (F.1.6),

(qn)i = Rni (qn)i = Rni x-space

(q'n)i = R'ni (q'n)i = R'ni x'-space . (F.5.4)

Now there are two affine connections, one for x-space and the other for x'-space,

Γcab ≡ (∂xc/∂ξi) (∂2ξi/∂xa∂xb) = Rci ∂a (∂ξi/∂xb) = Rci(∂aRbi) ∂a = ∂/∂xa

= [qc]i (∂a[qb]i) = qc • (∂aqb) (F.5.5)

Γ 'cab ≡ (∂x'c/∂ξi) (∂2ξi/∂x'a∂x'b) = R'ci ∂'a (∂ξi/∂x'b) = R'ci(∂'aR'bi) ∂'a = ∂/∂x'a

= [q'c]i (∂'a[q'b]i) = q'c • (∂'aq'b) . (F.5.6)

The expressions appearing here in (F.5.5) are just converted versions of (F.1.7) and (F.1.5).

319
 Appendix F: Affine Connection

F.6 Relations between Γ and Γ '

The claimed relations are the following, in the context of Picture A shown above,

Γ 'cab = Rcd Raα Rbβ Γdαβ + Rcα (∂'aRbα) // Weinberg p 100 (4.5.2) (F.6.1)

Γ 'cab = Rcd Raα Rbβ Γdαβ – Rbβ(∂'aRcβ) (F.6.2)

Γ 'cab = Rcd Raα Rbβ Γdαβ – Raα Rbβ (∂αRcβ) // Weinberg p 102 (4.5.8) (F.6.3)

The first terms are all the same, only the second terms vary. If the second term were not present, the
relation would state that Γdαβ transforms as a mixed rank-3 tensor in the usual manner, as in example
(7.10.1). Since the second term is present, Γdαβ is not a tensor (unless Rbα(x) = constant, in which case
the transformation F is linear so x' = F(x) = Rx as mentioned in Section 2.8 ) .

Proof of the first relation : We now make use of Picture D1 shown above. Start with the Γ ' definition
expression shown in (F.5.6),

Γ 'cab ≡ R'cn(∂'aR'bn) = (RR)cn ∂'a(RR)bn = RcdRdn ∂'a(RbβRβn) // R' = RR (F.5.3)

= RcdRdnRbβ(∂'aRβn) + Rcd(RdnRβn)( ∂'aRbβ) // RdnRβn = δdβ (7.6.4)

= RcdRdnRbβ([Raα∂α]Rβn) + Rcd(δdβ)( ∂'aRbβ) // ∂'a = Raα∂α in first term only

= RcdRaαRbβRdn(∂αRβn) + Rcβ (∂'aRbβ) // next use (F.5.5) for Rdn(∂αRβn)

= RcdRaαRbβ Γdαβ + Rcα (∂'aRbα) // Γdαβ ≡ Rdn(∂αRβn) QED (F.6.4)

Magically, all the R's have gone away.

The second term in the above relation can be written a different manner as follows. Consider,

0 = ∂'a(δcb) = ∂'a(RcαRbα) = Rcα (∂'aRbα) + Rbα (∂'aRcα)

⇒ Rcα (∂'aRbα) = – Rbα(∂'aRcα) = – Rbβ(∂'aRcβ)

= – Rbβ Raα(∂αRcβ) (F.6.5)

and this gives the other two relations stated above.

F.7 Statement and Proof of the Covariant Derivative Theorem

Many examples of this theorem will be given later. In this proof it is assumed that the tensor density of
interest is purely covariant (all indices "down"). In the next Section it will then be shown how to adjust
the theorem if one or more of the tensor density indices is "up".

320
 Appendix F: Affine Connection

Covariant Derivative Theorem: The covariant derivative Babc..x;α (as defined below) of a covariant
tensor Babc..x of rank n and weight W transforms as a covariant tensor density of rank n+1 and weight
W. (F.7.1)

The implication is that all the indices including α of Babc..x;α can be treated as ordinary tensor indices
with respect to raising, lowering, contraction, and so on. The first term below in Babc..x;α is the regular
derivative ∂α Babc..x, often written as Babc..x,α (comma, not semicolon), and this first term is not a
tensor. Only when all the "correction terms" are included does the object become a tensor.

The covariant derivative in x-space and then in x'-space is defined as follows: (Weinberg p 104 4.6.12)

Babc..x;α ≡ ∂α Babc..x – ΓnaαBnbc..x – ΓnbαBanc..x – .... – ΓnxαBabc..n // x-space

del a-term b-term x-term
+ [W/(2g)] (∂αg) Babc..x (F.7.2)

B'abc..x;α ≡ ∂'α B'abc..x – Γ 'naαB'nbc..x – Γ 'nbαB'anc..x – .... – Γ 'nxαB'abc..n // x'-space

del a-term b-term x-term
+ [W/(2g')] (∂'αg') B'abc..x (F.7.2)'

The two definitions are the same except everything is primed in x'-space (except constant weight W). This
is as one would expect if the x-space equation were a "true tensor equation" as discussed in Section 7.15
and were therefore "covariant". Although the Γ objects are not tensors themselves, the combination of
terms shown in the definition of Babc..x;α is a rank n+1 tensor (as will be demonstrated).

As was shown in (F.4.2), (1/2)(1/g)∂αg = Γκκα so the W terms could be written as W Γκκα Babc..x and
W Γ 'κκα B'abc..x, and this form is commonly seen in the literature on this subject.

A proof of the theorem must then show that Babc..x;α as defined above in fact transforms as a tensor
density of rank n+1 and weight W, which is to say, one must show that

B'abc..x;α = J-W Raa'Rbb'..... Rxx' Rαα' Ba'b'c'..x';α' (F.7.3)

or, changing abc..x→ABC..X and then a'b'c'.. → abc.. ,

B'ABC..X;α
= J-WRαα'{RAaRBb..... RXx} *
Babc..x';α' (F.7.4)

or, in gory detail,

321
 Appendix F: Affine Connection

∂'αB'ABC..X – Γ 'nAαB'nBC..X – Γ 'nBαB'AnC..X – ........... – Γ 'nXαB'ABC..n // LHS

del' a'-term b'-term x'-term
+ [W/(2g')] (∂'αg') B'ABC..X

= J-W Rαα'{RAaRBb..... RXx} * // RHS

{∂α'Babc..x – Γnaα'Bnbc..x – Γnbα'Banc..x – ........... – Γnxα'Babc..n
del a-term b-term x-term
+ ( [W/(2g)] (∂α'g) Babc..x } (F.7.5)
Proof:

There is probably a faster way to prove that (F.7.5) is valid, but we shall be content with a brute force
proof where we in essence evaluate both sides of (F.7.5) and show they are the same. In doing so, we
make use of identities obtained in previous Sections. The proof is carried out in five steps.

1. Expand the LHS del' term and show del'-del matches the RHS del term.

∂'αB'ABC..X = (Rαβ∂β)( J-W RAaRBb..... RXx Babc..x) del'

= Rαβ(∂β J-W) RAaRBb..... RXx Babc..x del'-J

+ Rαβ J-W (∂βRAa)RBb..... RXx Babc..x del'-a

+ Rαβ J-W RAa(∂βRBb)..... RXx Babc..x del'-b

...
+ Rαβ J-W RAaRBb. ... (∂βRXx) Babc..x del'-x

+ Rαβ J-W RAaRBb.............RXx (∂β Babc..x) del'-del (F.7.6)

We claim that this last del'-del term of the LHS matches the del term on the RHS:

del'-del LHS = Rαβ J-W {RAaRBb...........RXx} (∂β Babc..x) (F.7.7)

del RHS = J-W Rαα'{RAaRBb..... RXx} {∂α' Babc..x} (F.7.8)

Setting α' = β shows that these two terms indeed match.

2. Show that the a-related terms balance. The a'-term on the LHS of (F.7.5) is

– Γ 'nAαB'nBC..X . // a'-term of LHS (F.7.9)

The connection (F.6.3) between Γ ' and Γ reads (then shuffle indices as shown),

Γ 'cab = Rcd Raα Rbβ Γdαβ – Raα Rbβ (∂αRcβ) // α→κ, β→σ (F.6.3)
Γ 'cab = Rcd Raκ Rbσ Γdκσ – Raκ Rbσ (∂κRcσ) // c→n, a→A, b→α
Γ 'nAα = Rnd RAκ Rασ Γdκσ – RAκ Rασ (∂κRnσ) . (F.7.10)

322
 Appendix F: Affine Connection

Now insert into (F.7.9) the expression (F.7.10) for Γ 'nAα and the tensor transformation rule for B'nBC..X,

– { Rnd RAκ Rασ Γdκσ – RAκ Rασ (∂κRnσ)} { J-W Rnn'RBbRCc....Rxx Bn'bc..x } // a'-term (F.7.11)

and to this we must add the contribution del'-a shown in (F.7.6).

Meanwhile, the RHS a-term in (F.7.5) is this

J-W Rαα'{RAaRBb..... RXx}{– Γnaα'Bnbc..x} // now do n→n'

= – { J-W Rαα'{RAaRBb..... RXx}{Γn'aα'Bn'bc..x} . // a-term of RHS (F.7.12)

We want to show that

a'-term + del'-a = a-term

or
LHS a'-term (F.7.11) + LHS del'-a term (F.7.6) = RHS a-term (F.7.12)
or
– { Rnd RAκ Rασ Γdκσ – RAκ Rασ (∂κRnσ)} { J-W Rnn'RBbRCc....Rxx Bn'bc..x }
+ Rαβ J-W (∂βRAn')RBb..... RXx Bn'bc..x // a→n' , this is the del'-a term

= – J-W Rαα'{RAaRBb..... RXx}{Γn'aα'Bn'bc..x} ? (F.7.13)

We can see that the factors J-W RBbRCc....Rxx Bn'bc..x are common to both sides so they can be removed
to give a simpler relation which we must show to be valid:

– { Rnd RAκ Rασ Γdκσ – RAκ Rασ (∂κRnσ)} { Rnn' }

+ Rαβ (∂βRAn')
= – Rαα'{RAa }{Γn'aα'} . (F.7.14)

In the first term orthogonality (7.6.4) says Rnd Rnn' = δdn' which pins d to n' in that term only, so the
above becomes

– RAκ Rασ Γn'κσ + Rnn'RAκ Rασ (∂κRnσ) + Rαβ (∂βRAn') = – RασRAκ Γn'κσ . (F.7.15)

The first term on the left cancels the only term on the right. Then do β→σ in the third term to get,

RAκ Rασ {Rnn'(∂κRnσ)} + Rασ (∂σRAn') = 0 . (F.7.16)

Now cancel the common Rασ factor and use the symmetry (∂κRnσ) = (∂σRnκ) to get

(∂σRAn') = – Rnn'RAκ (∂σRnκ) . (F.7.17)

Now in this order do σ→a, A→d, n→e, n'→n, κ→m to get

323
 Appendix F: Affine Connection

(∂aRdn) = –Ren Rdm (∂aRem) . (F.7.18)

But this is seen to be the third identity of (F.2.1)! By reversing the above sequence of steps, one shows
that the three a-related terms in the above LHS = RHS equation balance:

a'-term + del'-a = a-term.

or
– { Rnd RAκ Rασ Γdκσ – RAκ Rασ (∂κRnσ)} { J-W Rnn'RBbRCc....Rxx Bn'bc..x }
+ Rαβ J-W (∂βRAn')RBb..... RXx Bn'bc..x // a→n' , this is the del'-a term

= – J-W Rαα'{RAaRBb..... RXx}{Γn'aα'Bn'bc..x} ? (F.7.13)

In reversing the sequence, one of course adds back the deleted common factors as well as the various
implied index sums. It seems clearer to state the proof this way rather than start with the last equality with
no justification and artificially thread backwards to the desired equation. This method is used as well in
the next section.

3. Show that the b-related terms balance. We have proven that (F.7.13) is valid, and it is true for any
tensor Bn'bc..x , since this is a common factor on both sides. Eq (F.7.13) is therefore still valid if we
replace the tensor Bn'bc..x by the different tensor Qn'bc..x ≡ Bbn'c..x. Tensor Q is tensor B with the
first two indices swapped. Make this replacement, and after doing so, make the free index swap A↔B
and summation index swap a↔b. One ends up then with this known-valid equation:

– { Rnd RBκ Rασ Γdκσ – RBκ Rασ (∂κRnσ)} { J-W Rnn'RAaRCc....Rxx Ban'c..x }
+ Rαβ J-W (∂βRBn')RAa..... RXx Ban'c..x

= – J-W Rαα'{RBbRAa..... RXx}{Γn'bα' Ban'c..x } . (F.7.19)

As the reader may suspect, the three terms in this last equation will have this interpretation with respect to
equation (F.7.5) :

b'-term + del'-b = b-term .

To verify this claim, we treat the terms one at a time.

b'-term: First state the equation (F.7.9) = (F.7.11) on the first line below, then get the second line doing
A↔B and dummy b↔ a :

– Γ 'nAαB'nBC..X = – { Rnd RAκ Rασ Γdκσ – RAκ Rασ (∂κRnσ)} { J-W Rnn'RBbRCc....Rxx Bn'bc..x }
– Γ 'nBαB'nAC..X = – { Rnd RBκ Rασ Γdκσ – RBκ Rασ (∂κRnσ)} { J-W Rnn'RAaRCc....Rxx Bn'ac..x }

Since this second equation is valid for any tensor Bn'ac..x, it is true for Qn'ac..x ≡ Ban'c..x where the
first two indices are swapped. This gives,

– Γ 'nBαB'AnC..X = – { Rnd RBκ Rασ Γdκσ – RBκ Rασ (∂κRnσ)} { J-W Rnn'RAaRCc....Rxx Ban'c..x }

324
 Appendix F: Affine Connection

The left side is the b'-term on the LHS of (F.7.5) and the right side is the first term in (F.7.19).

del'-b: The del'-b term in (F.7.6) is the first line below,

Rαβ J-W RAa(∂βRBb )..... RXx Babc..x // b→n

Rαβ J-W RAa(∂βRBn')..... RXx Ban'c..x

The second line matches the second term in (F.7.19).

b-term: The b-term on the RHS of (F.7.5) is

– J-W Rαα'{RAaRBb..... RXx}[ Γnbα'Banc..x] // n→n'

– J-W Rαα'{RAaRBb..... RXx}[ Γn'bα'Ban'c..x]

The second line matches the right side of (F.7.19).

We have thus shown that the b terms balance in equation (F.7.5), just the way the a terms balance.

4. Similarly, the c,d.....x terms match. Just repeat step 3 above for each extra index.

5. It remains to show that the three terms so far neglected in (F.7.5) match as well. These terms are

LHS: Rαβ(∂β J-W) RAaRBb..... RXx Babc..x // del'-J in (F.7.6)

+ [W/(2g')] (∂'αg') B'ABC..X // the LHS W term (F.7.20)

RHS: J-W Rαα'{RAaRBb..... RXx} [W/(2g)] (∂α'g) Babc..x // the RHS W term

That is, one must show that

Rαβ(∂β J-W) RAaRBb..... RXx Babc..x + [W/(2g')] (∂'αg') B'ABC..X

= J-W Rαα'{RAaRBb..... RXx} [W/(2g)] (∂α'g) Babc..x . (F.7.21)

Inserting the tensor transformation (F.7.4) of B'ABC..X in the second term gives

Rαβ(∂β J-W) RAaRBb..... RXx Babc..x + [W/(2g')] (∂'αg') J-W{RAaRBb..... RXx} * Babc..x

= J-W Rαα'{RAaRBb..... RXx} [W/(2g)] (∂α'g) Babc..x . (F.7.22)

Set β = α' in the first term to get

Rαα'(∂α' J-W) RAaRBb..... RXx Babc..x + [W/(2g')] (∂'αg') J-W{RAaRBb..... RXx} * Babc..x

= J-W Rαα'{RAaRBb..... RXx} [W/(2g)] (∂α'g) Babc..x . (F.7.23)

325
 Appendix F: Affine Connection

Next, remove the common factor RAaRBb..... RXx Babc..x (and associated implied sums) to get

Rαα' (∂α' J-W) + [W/(2g')] (∂'αg') J-W = J-W Rαα' [W/(2g)] (∂α'g) . (F.7.24)

Since ∂'α = Rαα'∂α'(covariant vector transformation) this becomes

Rαα' (∂α' J-W) + [W/(2g')] (Rαα'∂α'g') J-W = J-W Rαα' [W/(2g)] (∂α'g)
or
(∂'α J-W) + [W/(2g')] (∂'αg') J-W = J-W [W/(2g)] (∂'αg) . (F.7.25)

Move the second term to the RHS and do the left side derivative to get

(-W) J-W-1(∂'α J) = J-W (W/2) [ (∂'αg)/g – (∂'αg')/g' ] (F.7.26)

so it remains then to show that

J-1(∂'α J) = (1/2)[ (∂'αg')/g' - (∂'αg)/g ] . (F.7.27)

From (5.12.14) one has J2 = g'/g and J = (g'/g)1/2 so we need to show that

∂'α (g'/g)1/2 = (1/2) (g'/g)1/2[ (∂'αg')/g' - (∂'αg)/g ]

or
(1/2) (g'/g)-1/2 ∂'α (g'/g) = (1/2) (g'/g)1/2[ (∂'αg')/g' - (∂'αg)/g ]
or
∂'α (g'/g) = (g'/g) [ (∂'αg')/g' - (∂'αg)/g ] (F.7.28)

Evaluation of the left side of (F.7.28) gives

∂'α (g'/g) = [ g(∂'αg') - g'(∂'αg)]/g2 = (1/g) (∂'αg') - (g'/g2)(∂'αg) .

Evaluation of right side of (F.7.28) gives

(g'/g) [ (∂'αg')/g' - (∂'αg)/g ] = (1/g)(∂'αg') - (g'/g2 ) (∂'αg) ,

Since these agree, (F.7.27) is valid and therefore working backwards we find that (F.7.20) is valid.

This concludes our lengthy 5-step proof of the Covariant Derivative Theorem stated in (F.7.1).

F.8 Rules for raising any index on a covariant derivative of a covariant tensor density.

The Rules are stated in (F.8.9) and (F.8.11), but will only make sense after looking at the examples
presented here.

326
 Appendix F: Affine Connection

Consider the general form given in (F.7.2) for a covariant derivative

Babc..x;α ≡ ∂α Babc..x – ΓnaαBnbc..x – ΓnbαBanc..x – .... – ΓnxαBabc..n // x-space

del a-term b-term x-term
+ (W/2g) (∂αg) Babc..x . (F.8.1)

Notice that there are n indices on Babc..x and there are n corresponding terms on the RHS in addition to
the del and W terms. Each index of Babc..x thus has its own "correction term". What happens if one of
the indices (say b) on Babc..x is raised? To find out, apply gβb to both sides. The effect of doing this is
trivial for all terms except the del term and the b-term, since b is a regular tensor index on all such terms,
so we get

Baβc..x;α ≡ gβb ∂α Babc..x – ΓnaαBnβc..x – gβb ΓnbαBanc..x – .... – ΓnxαBaβc..n

del a-term b-term x-term
+ (W/2g) (∂αg) Baβc..x . (F.8.2)
The del term can be written

gβb (∂α Babc..x) = ∂α (gβb Babc..x) – (∂α gβb) Babc..x

= ∂α Baβc..x – (∂α gβb) Babc..x . (F.8.3)

The second term del term here can be combined with the b-term to give

del-2nd + b-term = – (∂α gβb) Babc..x – gβb ΓnbαBanc..x

= – (∂α gβn) Banc..x – gβb ΓnbαBanc..x // b→n in first term only
= [– (∂α gβn) – gβb Γnbα] Banc..x . (F.8.4)

The identity (F.3.1) with n→ i reads,

(∂cgab) = – [gai Γ bci + gbi Γaci]

or
[– (∂cgab) – gai Γ bci] = gbi Γaci // do c→α, b→n, a→β
[– (∂αgβn) – gβi Γ nαi] = gni Γβαi // then i→b on the left
βn βb n ni β
[– (∂αg ) – g Γ αb] = g Γ αi . (F.8.5)

Replacing [...] in (F.8.4) by the last result gives,

del-2nd + b-term = gni Γβαi Banc..x = Γβαi Baic..x = Γβαn Banc..x (F.8.6)

so it has been shown that

Baβc..x;α ≡ ∂α Baβc..x – ΓnaαBnβc..x + Γβαn Banc..x – ... – ΓnxαBaβc..n + W Γκκα Babc..x

new b term (F.8.7)

Here then is a comparison where β→b in the second line:

327
 Appendix F: Affine Connection

Babc..x;α ≡ ∂α Babc..x – ΓnaαBnbc..x – ΓnbαBanc..x – .... – ΓnxαBabc..n + W Γκκα Babc..x

Babc..x;α ≡ ∂α Babc..x – ΓnaαBnbc..x + Γbαn Banc..x – .... – ΓnxαBabc..n + W Γκκα Babc..x
del a-term b-term x-term W-term
(F.8.8)
This demonstrates the following rule (stated for q = b or any other index):

Rule for raising some non-last index q:

(1) In all terms, raise the q-position B index (in the q-term that index is called n)
(2) in the q correction term, make the replacement – Γnqα → + Γqnα ( = Γqαn ) (F.8.9)

We refer to a correction term like the a-term and x-term in (F.8.8) as a "covariant correction term", and a
correction term like the b-term as a "contravariant correction term". Thus, for an arbitrary up/down
placement of the abc..x indices, there will be a covariant correction term for each down index, and a
contravariant correction term for each up index. Here is another example where x is also taken up:

Babc..x;α ≡ ∂α Babc..x – ΓnaαBnbc..x + Γbαn Banc..x – .... + ΓxnαBabc..n + W Γκκα Babc..x

del a-term b-term x-term W-term
(F.8.10)
where the contravariant correction terms are marked in red.

Rule for raising the last index α: Raising the ;α index must be done "manually" so the first term will have
gαα'∂α = ∂α' and all remaining terms will have explicit gαα' factors. (F.8.11)

F.9 Examples of covariant derivative expressions

Example J = 0 (covariant derivative of a scalar B) // J is the rank of the B tensor

B;α = ∂α B covariant vector // = B,α

B;α = ∂α B contravariant vector // = B,α (F.9.1)

Example J=1: (covariant derivative of a vector B)

Ba;α = ∂α Ba – ΓnaαBn covariant rank-2 tensor // 2nd term is sym on a↔α (F.9.2)

Ba;α = ∂α Ba + Γaαn Bn mixed rank-2 tensor (F.9.3)

To obtain Ba;α, write (F.9.2) as Ba;β = ∂βBa – ΓnaβBn and apply gαβ to both sides. The result is then
Ba;α = ∂αBa – gαβΓnaβBn .Finlly, raise a and alter the correction term to get Ba;α = ∂αBa + gαβΓanβBn .

Ba;α = ∂α Ba – gαβΓnaβBn mixed rank-2 tensor (F.9.4)

Ba;α = ∂α Ba + gαβΓaβn Bn contravariant rank-2 tensor (F.9.5)

328
 Appendix F: Affine Connection

Example J=2: (covariant derivative of a rank-2 tensor)

Bab;α ≡ ∂α Bab – ΓnaαBnb – ΓnbαBan covariant rank-3 tensor (F.9.6)

Bab;α ≡ ∂α Bab + Γaαn Bnb – ΓnbαBan etc. (F.9.7)

Bab;α ≡ ∂α Bab – ΓnaαBnb + Γbαn Ban (F.9.8)

Bab;α ≡ ∂α Bab + Γaαn Bnb + Γbαn Ban . (F.9.9)

Again, application of gαβ would give expressions for the other four possibilities with ;α being "up". These
"other possibilities" are always present, but we shall no longer mention them in the following examples.

Example J=3: (covariant derivative of a rank-3 tensor)

Babc;α ≡ ∂α Babc – ΓnaαBnbc – ΓnbαBanc – ΓncαBabn (F.9.10)

Babc;α ≡ ∂α Babc + Γaαn Bnbc – ΓnbαBanc – ΓncαBabn (F.9.11)

...

Babc;α ≡ ∂α Babc + ΓanαBnbc + ΓbnαBanc + ΓcnαBabn (F.9.12)

Special J=2 application to the metric tensor:

gab;α ≡ ∂α gab – Γnaαgnb – Γnbαgan =0 // by identity (F.3.2) (F.9.13)

gab;α ≡ ∂α gab + Γaαn gnb – Γnbαgan = 0 + Γaαb – Γabα = 0 // gab = δab (F.9.14)

gab;α ≡ ∂α gab – Γnaαgnb + Γbαn gan = 0 – Γbaα + Γbαa = 0 (F.9.15)

gab;α ≡ ∂α gab + Γaαn gnb + Γbαn gan =0 // by identity (F.3.1) (F.9.16)

The middle lines use the fact (7.4.19) that gab = δab. Since gab;α is a tensor, knowing that any one of the
above vanishes implies that all four lines vanish! The net result is

gab;α = gab;α = gab;α = gab;α = 0 . // Weinberg p 105 (4.6.16,17,18) (F.9.17)

The covariant derivative of any form of the metric tensor vanishes. As Weinberg points one, one knows
that in a quasi-Cartesian x-space gab;α = 0 since gab = Gaaδa,b = constant and Γ = 0. Then in any x'-
space g'ab;α = 0 as well since g'ab;α = Raa' Rbb' Rαα' ga'b';α' .

329
 Appendix F: Affine Connection

Example J=2: (double covariant derivatives)

Consider again the J=1 examples from above

Ba;α = ∂α Ba – ΓnaαBn (F.9.2)

Ba;α = ∂α Ba + Γaαn Bn . (F.9.3)

This applies to any vector Ba. As the J=0 example shows, B;a and B;a are bona-fide vectors (covariant
and contravariant components of the same vector ) and therefore

B;a;α = ∂α B;a – ΓnaαB;n (F.9.18)

B;a;α = ∂α B;a + Γaαn B;n (F.9.19)

where we have simply inserted a semicolon in each term. In these equations B is a scalar.

Consider again the J=2 examples from above,

Bab;α = ∂α Bab – ΓnaαBnb – ΓnbαBan (F.9.6)

Bab;α = ∂α Bab + Γaαn Bnb – ΓnbαBan . (F.9.7)

This applies to any rank-2 tensors Bab or Bab. But Ba;b and Ba;b are bona-fide rank-2 tensors, and
therefore we can restate the above as,

Ba;b;α = ∂α Ba;b – ΓnaαBn;b – ΓnbαBa;n (F.9.20)

Ba;b;α = ∂α Ba;b + Γaαn Bn;b – ΓnbαBa;n (F.9.21)

In a similar manner one can derive expressions for triple covariant derivatives and beyond. For example

Ba;b;c;α = ∂α Ba;b;c – ΓnaαBn;b;c – ΓnbαBa;n;c – ΓncαBa;b;n . (F.9.22)

The next examples are for tensor densities so there will be a W term as shown in (F.8.8).

Example J = 0 (scalar density of weight W) // recall Γκκα = (2g)-1∂αg from (F.4.2)

B;α = ∂αB + W Γκκα B (F.9.23)

Example J = 1 (vector density of weight W)

Ba;α = ∂α Ba – ΓnaαBn + W Γκκα Ba covariant rank-2 tensor density

Ba;α = ∂α Ba + Γaαn Bn + W Γκκα Ba (F.9.24)

330
 Appendix F: Affine Connection

Example J = 2 (rank-2 tensor density of weight W)

Bab;α = ∂α Bab – ΓnaαBnb – ΓnbαBan + W Γκκα Bab covariant rank-3 tensor density (F.9.25)
Bab;α = ∂α Bab + Γaαn Bnb – n a κ a
Γ bαB n + W Γ κα B b (F.9.26)
Bab;α = ∂α Bab – ΓnaαBnb + Γbαn Ban + W Γκκα Bab (F.9.27)
Bab;α = ∂α Bab + Γaαn Bnb + Γbαn Ban + W Γκκα Bab . (F.9.28)

As noted in Example 2 below (D.2.3), adding a factor gW/2 to a tensor density of weight W neutralizes the
weight, and the result is a regular tensor. Here then are a few examples in which this is done. Since the
product is a tensor, there are no W correction terms.

Example J = 0 (covariant derivative of a scalar density B of weight W)

(gW/2B);α = ∂α(gW/2B) covariant vector (F.9.29)

(gW/2B);α = ∂α(gW/2B) contravariant vector (F.9.30)

Example J=1: (covariant derivative of a vector density B of weight W)

(gW/2Ba);α = ∂α (gW/2Ba) – gW/2 Γnaα Bn covariant rank-2 tensor (F.9.31)

(gW/2Ba);α = ∂α (gW/2Ba) + gW/2 Γaαn Bn mixed rank-2 tensor (F.9.32)

Example J=2: (covariant derivative of a tensor density B of weight W)

(gW/2Bab);α = ∂α (gW/2Bab) – Γnaα(gW/2Bnb) – Γnbα(gW/2Ban) covariant rank-3 tensor (F.9.33)

(gW/2Bab);α = ∂α (gW/2Bab) + Γaαn (gW/2Bnb) – Γnbα(gW/2Ban) etc. (F.9.34)
(gW/2Bab);α = ∂α (gW/2Bab) – Γnaα(gW/2Bnb) + b W/2
Γ αn (g Ba )n
(F.9.35)
(gW/2Bab);α = ∂α (gW/2Bab) + Γaαn (gW/2Bnb) + Γbαn (gW/2Ban) (F.9.36)

F.10 The Leibniz rule for the covariant derivative of the product of two tensor densities

If A and B are arbitrary tensor densities each with an arbitrary set of up and down indices and arbitrary
weight, then the claim of the Leibniz or product rule is this:

(A----B----);α ≡ A----;α B---- + A---- B----;α // Weinberg p 105 (4.6.14) (F.10.1)

Recall from the Covariant Derivative Theorem (F.7.1) that A---- and A----;α have the same weight, call it
WA. Similarly, B---- and B----;α have the same weight WB. According to the outer product rule (D.2.3),
both terms on the RHS above have weight WA+WB and therefore this sum is the weight of the LHS (A---
-B----);α as well.

Proof of (F.10.1): Start with the covariant derivative of a generic tensor A in the form of (F.8.8),

331
 Appendix F: Affine Connection

A----;α = A----,α + (A index correction terms) + WA Γκκα A----

B----;α = B----,α + (B index correction terms) + WB Γκκα B---- (F.10.2)

where --- is any number of up/down indices. The "index correction terms" are those Γ terms discussed
below (F.8.9). One can then write out the two terms on the RHS of (F.10.1) above as:

A----;α B---- = A----,α B---- + (A index correction terms) B---- + [WA Γκκα A---- ] B----

A---- B----;α = A---- B----,α + A---- (B index correction terms) + A---- [WB Γκκα B---- ] .
(F.10.3)
Meanwhile, the LHS of (F.10.1) can be written as

(A----B----);α = (A----B----),α + ( all index correction terms) + (WA+ WB) Γκκα (A----B----) .
(F.10.4)
Momentarily ignoring the index correction terms, it is clear that the other terms match between LHS and
RHS. The W terms match by visual inspection, while the regular derivative terms match due to the
"regular" Leibniz rule for the derivative of a product

(A----B----),α = A----,α B---- + A---- B----,α (F.10.5)

which is to say

∂α(A----B----) = (∂αA----)B---- + A---- (∂α B----) . (F.10.6)

Consider now the LHS terms called (all index correction terms) above. This set of terms can be
partitioned into two groups,

(all index correction terms) = (terms involving A indices) + (terms involving B indices) . (F.10.7)

Let us pause to look at a simple example where ICT means we just show the index correction terms,

(AabBcd);α|ICT = – Γnaα(AnbBcd) – Γnbα(AanBcd) + Γcαn (AabBnd) + Γdαn (AabBcn) . (F.10.8)

The correction terms can be reordered in this way

= { – Γnaα(Anb) – Γnbα(Aan) } Bcd + Aab { ΓcαnBnd + ΓdαnBcn }

= { index correction terms for Aab } Bcd + Aab { index correction terms for Bcd } . (F.10.9)

Just so, in the general case one has

(A----B----);α|ICT = ( all index correction terms)

= { index correction terms for A---- } B---- + A---- { index correction terms for B---- } (F.10.10)

and this then shows that the index correction terms on the two sides of (F.10.1) do in fact match. QED

332
 Appendix F: Affine Connection

Once the above product rule is verified, it is then easy to generalize just as for regular derivatives:

(A----B----C----);α ≡ A----;α B---- C---- + A---- B----;α C---- + A---- B---- C----;α (F.10.11)

Examples with two vectors:

(AaBb);α = Aa;αBb + AaBb;α

(AaBb);α = Aa;αBb + AaBb;α
(AaBb);α = Aa;αBb + AaBb;α
(AaBb);α = Aa;αBb + AaBb;α (F.10.12)

Example with a scalar function A and a vector B:

(ABb);α = A;αBb + ABb;α = A,αBb + ABb;α // A;α = A,α, (F.9.1) (F.10.13)

A more general example:

(AabcBde);α ≡ Aabc;α Bde + Aabc Bde;α (F.10.14)

Theorem: Any object X--- for which (X---);α = 0 can be "extracted" from a covariant derivative group
which contains X---. That is to say (F.10.15)

(A---- X--- B-----);α = X--- (A---- B-----);α

Proof: By the Leibniz Rule (F.10.1),

(A---- X--- B-----);α = (A---- B-----);α X--- + (A---- B-----) (X---);α = X--- (A---- B-----);

We shall now apply this theorem in a few examples.

Example 1: Any constant can be extracted from a group.

(π A---- B-----);α = π (A---- B-----);α // extract a constant (F.10.16)

(εijk A---- B-----);α = εijk (A---- B-----);α // εijk = element of perm. tensor = ±1 or 0 (F.10.17)

Example 2: Since by (F.7.19) gab;α = 0 one can always extract gab from a group. This applies to gab as
well. And of course it applies to gab = δab since δab is a constant.

(A---- B--a--);α = (gab A---- B--b--);α = gab(A---- B--b--);α (F.10.18)

333
 Appendix F: Affine Connection

This says that lowering (or raising) an index "commutes" with covariant differentiation -- one can lower
an index ignoring the fact that :α is sitting there. But we already know this must be true because we know
that the object (A----B--b--);α is a true tensor, and gab can lower any index on a true tensor.

Fact : (|g|s);α = 0, s = real This is |g| raised to any real power s (s is not an index) (F.10.19)

Proof: Recall from (D.2.3a) that

|g'|-W/2 = J-W |g|-W/2 . // no R factors since scalar (D.2.3a)

Setting s = -W/2 gives

|g'|s = J2s |g|s

so |g'|s transforms as a scalar density of weight W = -2s. Recall next the derivative rule for a scalar
density of weight W,

B;α = ∂αB + W Γκκα B (F.9.23)

which we apply with B = |g|s to find,

(|g|s);α = ∂α|g|s -2s Γκκα |g|s

= ∂α|g|s -2s { (1/2)(1/|g|)∂α|g| } |g|s // (F.4.2) for Γκκα

= s|g|s-1 ∂α|g| - s |g|s-1∂α|g|

=0. QED

Example 3: Since by (F.10.19) (|g|s);α = 0, |g|s may be extracted from a covariant derivative group,

( |g|s A---- B-----);α = |g|s (A---- B-----);α (F.10.20)

334
 Appendix G: Expansion of (∇v)

Appendix G: Expansion of (∇v) in curvilinear coordinates (v = vector)

G.1 Continuum Mechanics motivation

This Appendix assumes the usual curvilinear coordinates context, Picture B

(G.1.1)

Although the polyadic notation is regarded as archaic by some writers (eg, Wolfram), it is well embedded
into the literature of continuum mechanics, a field awash in tensors.
In the literature one sometimes sees, in Cartesian coordinates,

(∇A)ij ≡ ∂jAi ≡ Ai,j (G.1.2)

where the indices are the reverse of the normal dyadic definition of (E.4.1),

(BA)ij ≡ BiAj . (G.1.3)

For example, in continuum mechanics one encounters the so-called convective or material derivative of
an arbitrary vector field A(x,t) in the Eulerian or spatial "view" of the motion of a blob of continuous
matter,

DAi/Dt = ∂tAi + ∇Ai • v = ∂tAi + (∂jAi) vj = ∂tAi + (∇A)ij vj = ∂tAi + [(∇A) v]i

⇒ DA/Dt = ∂tA + (∇A) v // = ∂tA + (v•∇)A = (∂t + v•∇) A = D/Dt (A). (G.1.4)

DAi/Dt is a historical notation for the total derivative dAi(x,t)/dt. Here v(x,t) is the velocity field of the
moving matter blob. An example is acceleration a, where A = v,

a = Dv/Dt = ∂tv + (∇v) v // = ∂tv + (v•∇)v = (∂t + v•∇) v = D/Dt (v). (G.1.5)

see Lai p 76 (3.4.3) and p 78 (3.4.8). The object (∇v), called the velocity gradient, is of great interest in
fluid mechanics. Correspondingly, the object (∇u) is of great interest in the theory of elastic solids, where
u is the displacement field.

The equation (G.1.4) written with proper index positions,

DAi/Dt = ∂tAi + (∇A)ij vj (G.1.4)

335
 Appendix G: Expansion of (∇v)

becomes a true tensor equation if we make the assumption shown in the next Section that (∇A)ij ≡ Ai;j .
In this case, we know from the covariance theorem exemplified in (E.9.13) that, for orthogonal
coordinates, this equation will appear in terms of ^en-expanded coefficients (like vj = h'jvj) as follows:

DA'i/Dt = ∂tA'i + (∇A)'ij v'j (G.1.6)

^
where (∇A)'ij = [(∇A)(e)]ij. To actually use equation (G.1.6) in some system of orthogonal curvilinear
^
coordinates, one has to compute the coefficients [(∇A)(e)]ij. This is done below in Section G.5 generally
and Section G.6 for a few specific coordinate systems (where A is called v) .

G.2 Expansion of ∇v on ei⊗ej by Method 1: Use the fact that vb;a is a tensor.

The covariant derivative vb;a is discussed in Sections F.7, F.8 and F.9. We shall define a true tensor
object (∇v) as vb;a so that, according to (F.9.2),

(∇v)ba ≡ vb;a = [vb,a – Γ cab vc] // vb,a ≡ ∂avb x-space

(∇v)'ba ≡ v'b;a = [v'b,a – Γ 'cab v'c] . // v'b,a ≡ ∂'av'b x'-space (G.2.1)

If x-space is Cartesian, then Γ = 0 as in (F.4.16), in which case

(∇v)ba = vb,a = ∂avb (G.2.2)

so (∇v)ba = vb;a aligns with our dyadic (∇v) object (E.4.4) in Cartesian space.
As shown in (E.2.14) one can expand the rank-2 tensor vb;a in either of these ways,

∇v = Σij vi;j ui⊗uj vi;j = [vi,j – Γcij vc] = vi,j = ∂jvi

∇v = Σij v'i;j ei⊗ej v'i;j = [v'i,j – Γ ' cij v'c] (G.2.3)

where the ui are the Cartesian basis vectors in x-space, while ei are the reciprocal base vectors.
According to (F.5.6) with Fig (F.5.2), the affine connection Γ ' in x'-space is given by Γ 'cab =
R'cn(∂'aR'bn). When x-space is Cartesian, R = 1 and then R' = R (F.5.3), so

Γ 'cab = Rci(∂'aRbi)
Γ 'cij = Rck(∂'jRik) . (G.2.4)

Inserting this into the second expansion of (G.2.3) gives

∇v = Σij v'i;j ei⊗ej v'i;j = [(∂'jv'i) – Rck(∂'jRik) v'c]

or
∇v = Σij [(∇v)(e)]ij ei⊗ej [(∇v)(e)]ij = [(∂'jv'i) – Rck(∂'jRik) v'c] . (G.2.5)

336
 Appendix G: Expansion of (∇v)

Note that the expression shown contains only x'-space coordinates and objects. The (e) superscript tells
us that this matrix element of operator (∇v) is taken in the en basis, see Section E.7:

[(∇v)(e)]ij = <ei|∇v| ej> = (ei)T (∇v) ej = ei • (∇v) ej = ei • (∇v) • ej (G.2.6)

bra-ket matrix dot of vectors dyadic

Recall now the third identity from (F.2.1),

(∂aRdn) = – Ren Rdm (∂aRem) 3 . (F.2.1)

This was derived for Picture C1 of (F.1.1) with x on the left and ξ on the right. Adjusting this to Picture B
of (G.1.1) with x' on the left and x on the right gives the first line below, and then an index shuffle gives
the second line,

(∂'aRdn) = – Ren Rdm (∂'aRem) // a→j, d→i, n→k

(∂'jRik) = – Rek Rim (∂'jRem) . (G.2.7)

Then,

Rck(∂'jRik) = – Rck Rek Rim (∂'jRem) = – δce Rim (∂'jRem) = –Rim(∂'jRcm) (G.2.8)

so that (G.2.5) can be written in this way,

∇v = Σij [(∇v)(e)]ij ei⊗ej [(∇v)(e)]ij = v'i;j = [(∂'jv'i) + Rim(∂'jRcm) v'c] . (G.2.9)

We shall use this second form for [(∇v)(e)]ij below.

Footnote: A variation of the above development would be to start this way, using the expansion of
(E.2.12) on the left, and using (F.9.3) on the right,

∇v = Σij vi;j ui⊗uj vi;j = [vi,j + Γijc vc] = vi,j = ∂jvi

∇v = Σij v'i;j ei⊗ej v'i;j = [v'i,j + Γ ' ijc v'c] (G.2.10)

Eq (G.2.4) says Γ 'cab = Rci(∂'aRbi), but (F.1.10) gives the alternate form shown on the first line below,
and then an index shuffle gives the second line,

Γ 'cab = – Rbi (∂'aRci) // i→m, then c→i, a→j, b→ c

Γ ' ijc = – Rcm (∂'jRim) .

Then (G.2.10) says

∇v = Σij [(∇v)(e)]ij ei⊗ej [(∇v)(e)]ij = v'i;j = [(∂'jv'i) – Rcm (∂'jRim) v'c] (G.2.11)

which is similar to (G.2.5) but applies to the down-tilt index configuration.

337
 Appendix G: Expansion of (∇v)

G.3 Expansion of ∇v on ei⊗ej by Method 2: Use brute force.

Method 1 is perhaps elegant in that it makes use of tensor transformations and the affine connection Γ.
But a simple brute force method is really just as simple and does not require knowledge of Γ and
covariant differentiation. Instead of using the ei⊗ej notation for basis vectors, here we use the alternate
notation ei(ej)T which works for rank-2 tensor expansions. Recall that ei(ej)T is an NxN matrix as
illustrated in (E.5.3) for N = 2.

In this brute force method, start with the Cartesian-space expansion,

(∇v) = Σcd(∇v)dc uducT = Σcd(∂cvd) uducT . // note that ud = ud in Cartesian space (G.3.1)

This expansion is the second line of (E.2.14) using the matrix notation (E.5.3), ud ⊗ uc = uducT .
To express things in x'-coordinates, first write

∂cvd = (Ric∂'i)( Rjdv'j) = Ric [(∂'iRjd) v'j + Rjd (∂'iv'j)] . (G.3.2)

The next step is to express the matrix uducT as a linear combination of eeefT. We start with the
orthogonality rule # 1 of (7.6.4) .

δni = Rbn Rbi // implied sum on b, as usual (7.6.4)

We know that

(en)i = Rni (7.18.1)

(un)i = δni (7.18.3)

so the above orthogonality rule says:

(un)i = Rbn (eb)i ⇒ un = Rbn eb

or ud = Red ee . (G.3.3)

Taking the transpose of both sides (sum of column vectors to sum of row vectors),

uTd = Red eTe ⇒ ucT = Rfc efT . (G.3.4)

Combining these last two results gives

uducT = Red Rfc ee efT = Red Rfc ee efT . (G.3.5)

The expansion (G.3.1) of (∇v) may then be written,

338
 Appendix G: Expansion of (∇v)

(∇v) = (∇v)dc uducT = (∂cvd) uducT

= Ric [(∂'iRjd) v'j + Rjd (∂'iv'j)] { Red Rfc ee efT } // (G.3.2) and (G.3.5)

= (Rfc Ric)[Red (∂'iRjd) v'j + (Red Rjd)(∂'iv'j)] ee efT

= δfi [Red (∂'iRjd) v'j + δej (∂'iv'j)] ee efT // orthogonality twice

= [Red (∂'fRjd) v'j + (∂'fv'e)] ee efT

= Σef [(∇v)(e)]ef ee efT (G.3.6)

where

[(∇v)(e)]ef = (∂'fv'e) + Red (∂'fRjd) v'j (G.3.7)

or
[(∇v)(e)]ij = (∂'jv'i) + Rim (∂'jRjm) v'j (G.3.8)

This agrees with (G.2.9) obtained by the previous method.

G.4 Expansion on ei⊗ej and ^

e i⊗^
ej

Reversing the index tilts in (G.2.5) [as justified in (E.2.12)] one gets

∇v = Σij v'i;j ei⊗ej

where v'i;j = g'iag'jb v'a;b = g'iag'jb [(∂'bv'a) + Ram(∂'bRcm) v'c] . // (G.2.9) (G.4.1)

Then since ei = h'i^

e i one gets yet another expansion (more generally see Section E.8 ),

^
∇v = Σij (v'i;j h'i h'j) ^
e i⊗^
e j = Σij [(∇v)(e)]ij ^
e i⊗^
ej

^
where [(∇v)(e)]ij = h'i h'j v'i;j = h'i h'j g'iag'jb [(∂'bv'a) + Ram(∂'bRcm) v'c] . (G.4.2)

Here is a summary of results so far:

∇v = Σij [(∇v)(e)]ij ei⊗ej [(∇v)(e)]ij = [(∂'jv'i) + Rim(∂'jRcm) v'c] (G.2.9)

∇v = Σij [(∇v)(e)]ij ei⊗ej [(∇v)(e)]ij = [(∂'jv'i) – Rcm (∂'jRim) v'c] . (G.2.11)

∇v = Σij [(∇v)(e)]ij ei⊗ej [(∇v)(e)]ij = g'iag'jb [(∂'bv'a) + Ram(∂'bRcm) v'c] (G.4.1)

^ ^
∇v = Σij [(∇v)(e)]ij ^
e i⊗^
ej [(∇v)(e)]ij = h' h'j g'iag'jb [(∂'bv'a) + Ram(∂'bRcm) v'c] (G.4.2)
(G.4.3)

339
 Appendix G: Expansion of (∇v)

One could replace v'a = g'adv'd in any of the above results. For example, the last object becomes
^
[(∇v)(e)]ij = h'i h'j g'iag'jb [(∂'b[g'adv'd]) + Ram(∂'bRcm) (g'cdv'd)] . (G.4.4)

Another choice is to use the v'n components of v obtained when expanding v on the ^
en ,

v = Σn v'n en = Σn v'n (h'n^ e n = Σn v'n ^

e n) = Σn (h'n v'n) ^ en ⇒ v'n ≡ h'n v'n (G.4.5)

so one then replaces v'n = h'n-1v'n. The same object above then becomes
^
[(∇v)(e)]ij = h'i h'j g'iag'jb [(∂'b[g'ad h'd-1v'd]) + Ram(∂'bRcm) (g'cd h'd-1v'd)] . (G.4.6)

As discussed in Section 14.7, the components v'n are convenient since they all have the same dimensions.
Moreover, when a specific curvilinear system is selected, one can dispense with the unpleasant font used
in v'n and just write v'n = vx'(n). For example, in spherical coordinates r,θ,φ :

v'1 = vr v'2 = vθ v'3 = vφ v = Σn v'n ^ r + vθ ^

e n = v r^ ^ .
θ + vφ φ (G.4.7)

In Chapter 14 the scripted variables have no primes because Picture (14.1.1) is being used instead of
Picture B.

G.5 Orthogonal coordinate systems

In this case g'ab = h'a2δa,b and g'ab = h'a-2δa,b and things simplify. The four expansions (G.4.3) become
(there is no change in the first two expansions)

∇v = Σij [(∇v)(e)]ij ei⊗ej [(∇v)(e)]ij = [(∂'jv'i) + Rim(∂'jRcm) v'c]

∇v = Σij [(∇v)(e)]ij ei⊗ej [(∇v)(e)]ij = [(∂'jv'i) – Rcm (∂'jRim) v'c] .

∇v = Σij [(∇v)(e)]ij ei⊗ej [(∇v)(e)]ij = h'i-2 h'j-2 [(∂'jv'i) + Rim(∂'jRcm) v'c]

^ ^
∇v = Σij [(∇v)(e)]ij ^
e i⊗^
ej [(∇v)(e)]ij = h'i-1 h'j-1 [(∂'jv'i) + Rim(∂'jRcm) v'c] (G.5.1)

^
The (G.4.6) version of [(∇v)(e)]ij becomes,

^
Pij ≡ [(∇v)(e)]ij = h'i-1 h'j-1 [(∂'j[h'iv'i]) + Rim(∂'jRcm) (h'cv'c)]

= h'i-1 h'j-1 [(∂'j[h'iv'i]) + Rid(∂'jRbd) (h'bv'b)] // m→d,c→b

= h'i-1 h'j-1 [ (∂'jh'i) v'i + h'i(∂'jv'i) + h'i2Rid(∂'jRbd) (h'bv'b)] (G.5.2)

T2 T3 T1

where Rid = g'iaRab gbd = h'i2Rid is used to put all R's into their down-tilt form.

340
 Appendix G: Expansion of (∇v)

G.6 Maple evaluation of (∇v) in several coordinate systems

This object Pij shown above can be computed in Maple by a simple program which is easily modified for
other orthogonal curvilinear systems. The first task is to obtain the R matrix from the inverse
transformation equations x = F-1(x'), where we assume spherical coordinates as an example. In this code,
N = 3 and p means prime, as in hp = h'.

(G.6.1)

Then one needs the scale factors hn' from the metric tensor ḡ' = STS in (5.7.9) [developmental notation],

341
 Appendix G: Expansion of (∇v)

(G.6.2)

The terms T1,T2 and T3 shown in (G.5.2) are then entered,

^
Pij ≡ [(∇v)(e)]ij = polh'i
-1
h'j-1 [ (∂'jh'i) v'i + h'i(∂'jv'i) + h'i2Rid(∂'jRbd) (h'bv'b)] (G.5.2)
T2 T3 T1

(G.6.3)

The terms are then added and simplified and out pops the result,

(G.6.4)
Below are some sample results (including the above):

342
 Appendix G: Expansion of (∇v)

^
(∇v) = Σij Pij ^
ei ^
e jT Pij = [(∇v)(e)]ij (G.6.5)

• Pij in polar coordinates (where 1,2 = r,θ) :

// agrees with Lai p 57 (2.33.23) (G.6.6)

• Pij in cylindrical coordinates (where 1,2,3 = r,θ,z) :

// agrees with Lai p 60 (2.34.5) (G.6.7)

The polar coordinates results are seen to be the upper-left 2x2 piece of the cylindrical results.

• Pij in spherical coordinates (where 1,2,3 = r,θ,φ) :

// agrees with Lai p 64 (2.35.25) (G.6.8)

By entering the usual inverse transformation equations x' = F-1(x), and with suitable small alterations, the
above Maple code can compute ∇v in any orthogonal curvilinear coordinate system in any number of
dimensions N.

343
 Appendix H: Expansion of div(T)

Appendix H: Expansion of div(T) in curvilinear coordinates (T = rank-2 tensor)

H.1 Introduction

The object of attention in this Appendix, divT, is expressed this way in Cartesian coordinates,

(divT)i = ∂jTij , (H.1.1)

where Tij is a rank-2 tensor. One can regard the above equation as describing the normal divergence of
the "vector" which forms the ith row of the matrix Tij. In general (that is to say, under general
transformations F), the rows of Tij are not tensorial vectors, and that is why (divT)i is not a tensorial
scalar. Nor, in fact, are the (divT)i as defined above the components of a tensorial vector! As shown in
Section H.3 below, divT will be redefined as a true tensorial vector which equals ∂jTij in Cartesian
coordinates.

H.2 Continuum Mechanics motivation

The vector divT arises for example when Newton's 2nd Law F = ma is applied to a particle of continuous
matter, in which case this law is known as Cauchy's Equation of Motion,

divT + ρB = ρa // Lai p 169 (4.7.4) (H.2.1)

In this equation T is known as the Cauchy stress tensor, ρ is the mass density of the particle, B is any
action-at-a-distance force (body force) per unit mass (such as gravity), and of course a is the acceleration
of the particle.

Under rotations and translations the quantity (divT)i = ∂jTij is a tensorial vector, ρ is a tensorial scalar,
and a and B are tensorial vectors. As one would expect, divT + ρB = ρa is covariant in the sense of
Section 7.15 under these kinds of transformations.

More generally, if (divT)a is taken to mean Tab;b as shown in the next Section. then (H.2.1) is a true
tensor equation under any transformation x' = F(x). Then as shown in (E.9.13), equation (H.2.1) stated in
terms of ^en-expanded coefficients takes the form

(divT)' + ρB' = ρa' (divT)'i = [(divT)(e)]i (H.2.2)

where en are the tangent base vectors for some system of orthogonal curvilinear coordinates. One again,
in order to use this equation, one must compute the coefficients [(divT)(e)]i and this task is done below.
The result is rather complicated, as shown in (H.5.7). For spherical coordinates, the three components of
(divT)'i = [(divT)(e)]i are displayed in (H.6.3), where each component is seen to have seven terms.

344
 Appendix H: Expansion of div(T)

H.3 Expansion of divT on en by Method 1: Use fact that Tab;α is a tensor.

As shown in (F.9.9), the following object transforms as a rank-3 tensor,

Tab;α ≡ ∂αTab + Γaαn Tnb + Γbαn Tan = ∂α Tab // x-space [g=1, Γ = 0 (F.4.16) ]
T 'ab;α ≡ ∂'αT 'ab + Γ 'aαn T 'nb + Γ 'bαn T 'an // x'-space (H.3.1)

Contracting b with α yields the following tensorial vector equations,

(divT)a = Tab;b = ∂bTab // x-space , Γ = 0

(divT)'a = T 'ab;b ≡ ∂'bT 'ab + Γ 'abn T 'nb + Γ 'bbn T 'an // x'-space (H.3.2)

Note that the Cartesian space statement (divT)a = ∂bTab is obtained, as noted in the opening comments
above. As in (E.2.10), the vector divT can be expanded as

divT = Σa(divT)'a ea (divT)'a = [(divT)(e)]a // divT in the en basis (H.3.3)

Recall from (F.1.10) that

Γcab = Rci (∂aRbi) // for Picture C1 in (F.1.1)

Γcab = – Rbi (∂aRci) . (F.1.10)

Adjusting from Picture C1 to Picture A,

the last equation above becomes the first one below. We then shuffle indices to get the second line, and
contract a with b to get the third:

Γ 'cab = – Rbi (∂'aRci) // do b→n then a→b then c→a

Γ 'abn = – Rni (∂'bRai) // next, contract indices a and b
Γ 'bbn = – Rni (∂'bRbi) . (H.3.4)

Installing the last two lines into the second line of (H.3.2) then gives

(divT)'a = ∂'bT 'ab + Γ 'abn T 'nb + Γ 'bbn T 'an

= ∂'bT 'ab – Rni (∂'bRai) T 'nb – Rni (∂'bRbi)T 'an . (H.3.5)

The conclusion is that

345
 Appendix H: Expansion of div(T)

divT = Σa[(divT)(e)]a ea (H.3.6)

[(divT)(e)]a = ∂'b T 'ab – Rni(∂'bRai) T 'nb – Rni (∂'bRbi)T 'an // sum on n and b

The vector divT = Σa(∂bTab)ua has thus been expressed in terms of x'-space coordinates and objects,
and as usual the en are the tangent base vectors in x-space.

H.4 Expansion of divT on en by Method 2: Use brute force.

Start with the known expansion of divT in Cartesian x-space, as in (E.2.10),

divT = Σi [divT]i ui = Σi (∂jTij) ui . (H.4.1)

Compute

(∂jTij) = (Raj∂'a)( Rb'i Rc'j T 'b'c')

= Raj Rb'i Rc'j (∂'a T 'b'c') + Raj Rb'i (∂'a Rc'j) T 'b'c' + Raj Rc'j (∂'a Rb'i) T 'b'c'

= (Raj Rc'j) Rb'i (∂'a T 'b'c') + Raj Rb'i (∂'a Rc'j) T 'b'c' + (Raj Rc'j) (∂'a Rb'i) T 'b'c'

= δac' Rb'i (∂'a T 'b'c') + Raj Rb'i (∂'a Rc'j) T 'b'c' + δac' (∂'a Rb'i) T 'b'c'

= Rb'i (∂'a T 'b'a) + Raj Rb'i (∂'a Rc'j) T 'b'c' + (∂'a Rb'i) T 'b'a . (H.4.2)

In (G.3.3) it was shown that un = ΣbRbn eb. This is valid with indices tilted the other way, as the reader
can show mimicking the derivation of (G.3.3) or by raising and lowering labels using (7.18.1) en = g'ni ei
and (7.18.3) un = gni ui and (7.5.9) Rab = g'aa'Ra'b' gb'b . The result is

ui = Σe Rni en . (H.4.3)

Inserting (H.4.2) and (H.4.3) into (H.4.1) gives, twice using orthogonality rules (7.6.4),

divT = Σi (∂jTij) ui

= { (Rni Rb'i) (∂'a T 'b'a) + Raj (Rni Rb'i) (∂'a Rc'j) T 'b'c' + Rni (∂'a Rb'i) T 'b'a} en

= { δnb' (∂'a T 'b'a) + Raj δnb' (∂'a Rc'j) T 'b'c' + Rni (∂'a Rb'i) T 'b'a} en

= { (∂'a T 'na) + Raj (∂'a Rc'j) T 'nc' + Rni (∂'a Rb'i) T 'b'a} en

= Σn[(divT)(e)]n en (H.4.4)

where

346
 Appendix H: Expansion of div(T)

[(divT)(e)]n = (∂'a T 'na) + Raj (∂'a Rc'j) T 'nc' + Rni (∂'a Rb'i) T 'b'a . (H.4.5)

The third identity in (F.2.1) says

(∂aRdn) = – Ren Rdm (∂aRem) 3 . // for Picture C1 of (F.1.1) (F.2.1)

Adjusted to Picture C1' of (F.1.4) this becomes the first line below, then the next two lines are obtained
by index shuffles,

(∂'aRdn) = – Ren Rdm (∂'aRem) // do d→c', n→j

(∂'aRc'j) = – Rej Rc'm (∂'aRem) // do j → i
(∂'aRb'i) = – Rei Rb'm (∂'aRem) . (H.4.6)

Inserting these last two lines into (H.4.5) gives,

[(divT)(e)]n = (∂'a T 'na) + Raj (∂'a Rc'j) T 'nc' + Rni (∂'a Rb'i) T 'b'a (H.4.5)

= (∂'a T 'na) – Raj Rej Rc'm (∂'aRem) T 'nc' – Rni Rei Rb'm (∂'aRem) T 'b'a

= (∂'a T 'na) – δaeRc'm (∂'aRem) T 'nc' – δne Rb'm (∂'aRem) T 'b'a // orthog.

= (∂'a T 'na) – Rc'm (∂'aRam) T 'nc' – Rb'm (∂'aRnm) T 'b'a . (H.4.7)

Reverse the order of the last two terms,

[(divT)(e)]n = (∂'a T 'na) – Rb'm (∂'aRnm) T 'b'a – Rc'm (∂'aRam) T 'nc' . (H.4.8)

Replace summation indices m→i, a→b,

= (∂'b T 'nb) – Rb'i (∂'bRni) T 'b'b – Rc'i (∂'bRbi) T 'nc' . (H.4.9)

Finally, change n→a and then c'→ n and b'→n,

[(divT)(e)]a = (∂'b T 'ab) – Rni (∂'bRai) T 'nb – Rni (∂'bRbi) T 'an . // sum on n and b (H.4.10)

This is seen to match the result (H.3.6) of the previous Section. The brute force method is in fact not too
bad and requires no explicit use of the affine connection Γ.

Technical Note: In the above expression one can write for example Rni(∂'bRai) = g'nmRmi(∂'bRai). Then
using the fact that RmiRai = g'ma one finds Rmi(∂'bRai) + Rai(∂'bRmi) = (∂'bg'ma) which then allows the
replacement Rni(∂'bRai) = g'nm (∂'b g'ma) – g'nm Rai(∂'bRmi). This kind of transformation leads to an
alternate form for divT, still with all down-tilt R matrices, but the index structure is different. This other
form appears when one does the "brute force" method starting with (∂jTij) instead of (∂jTij). Of course
in Cartesian space these two objects must be the same.

347
 Appendix H: Expansion of div(T)

H.5 Adjustment for T expanded on (^ei⊗e^j) and divT expanded on ^ea

In the above, it has been assumed that Tij is a rank-2 tensor so that, in the notation of (E.2.11),

T = ΣijTij(ui⊗uj) = ΣijT 'ij(ei⊗ej) . (H.5.1)

If one is interested in an expansion of T on the unit vectors ^

e i where ei = h'i ^
e i this becomes

^
T = Σij[T 'ijh'ih'j] (^
e i⊗^
e j) = Σij[T(e)]ij (^
e i ⊗^
e j) (H.5.2)

One then has

^
[T(e)]ij = h'ih'jT 'ij ⇒ T 'ij = h'i-1 h'j-1 [T(e^)]ij . (H.5.3)

If one is interested in this form of the T matrix elements, then one is likely also interested in this
expansion for divT,

^
divT = Σa[(divT)(e)]a ea = Σn { [(divT)(e)]a h'a } ^
e a ≡ [(divT)(e)]a ^
ea (H.5.4)

where then

^
[(divT)(e)]a = h'a[(divT)(e)]a

= h'a [(∂'b T 'ab) – Rni (∂'bRai) T 'nb – Rni (∂'bRbi) T 'an] // from (H.4.10)

^ ^
= h'a * { ∂'b (h'a-1 h'b-1 [T(e)]ab) – Rni (∂'bRai) h'n-1 h'b-1 [T(e)]nb
^
– Rni (∂'bRbi) h'a-1 h'n-1 [T(e)]an } // from (H.5.3)

^ ^
= h'a * { ∂'b (h'a-1 h'b-1) [T(e)]ab – Rni (∂'bRai) h'n-1 h'b-1 [T(e)]nb // ∂(xy) = dx y + x dy
-1 -1 ^) ab
(e i b -1 -1 ^) an
(e
+ h'a h'b (∂'b [T ] ) – Rn (∂'bR i) h'a h'n [T ] } .
(H.5.5)
Now

∂'b (h'a-1 h'b-1) = ∂'b(h'a h'b)-1 = – (h'a h'b)-2 ∂'b(h'a h'b) = – h'a-2 h'b-2 ∂'b(h'a h'b)

^
so the above sequence for [(divT)(e)]a continues,

^ ^
= h'a * { – h'a-2 h'b-2 ∂'b(h'a h'b) [T(e)]ab – Rni (∂'bRai) h'n-1 h'b-1 [T(e)]nb
^ ^
+ h'a-1 h'b-1 (∂'b [T(e)]ab) – Rni (∂'bRbi) h'a-1 h'n-1 [T(e)]an }

^ ^
= { – h'a-1 h'b-2 ∂'b(h'a h'b) [T(e)]ab – h'a h'n-1 h'b-1 g'nm Rmi (∂'bRai) [T(e)]nb
^ ^
+ h'b-1 (∂'b [T(e)]ab) – h'n-1 g'nm Rmi (∂'bRbi) [T(e)]an } (H.5.6)

348
 Appendix H: Expansion of div(T)

where recall that Rni = Rni since x-space is Cartesian. In the last form only the down-tilt R matrix
appears which simplifies calculation with Maple. This and all other results above are valid for general
curvilinear coordinates, orthogonal as well as non-orthogonal.

At this point we specialize to orthogonal systems, so (H.5.6) becomes

T1 T3
^) a
(e -1 -2 ^) ab
(e ^
[(divT) ] = { – h'a h'b ∂'b(h'a h'b) [T ] – h'a h'b h'n R i (∂'bRai) [T(e)]nb
-1 n
^ ^
+ h'b-1 (∂'b [T(e)]ab) – h'n Rni (∂'bRbi) [T(e)]an }
T2 T4 (H.5.7)

Even for an orthogonal curvilinear coordinate system, the form of this tensor divergence is amazingly
complicated. Here it is expressed in terms of the down-tilt R matrix and the scale factors, and as usual
repeated indices are summed.

H.6 Maple: divT in cylindrical and spherical coordinates

^
The above object [(divT)(e)]a can be evaluated by Maple code very similar to that shown in Appendix G
for (∇v). The main difference is the set of entry lines for the terms,

(H.6.1)

349
 Appendix H: Expansion of div(T)

^
Here are some results for [(divT)(e)]a = "divTa" :

• cylindrical coordinates (where 1,2,3 = r,θ,z) :

// agrees with Lai p 60 (2.34.8,9,10)

(H.6.2)

• For polar coordinates (divT)r,θ and are given by the first two lines above with the ∂z terms set to 0.
These polar results then agree with Lai p 58 (2.33.32,33).

• spherical coordinates (where 1,2,3 = r,θ,φ) :

(H.6.3)

The expressions above agree with Lai p 65 (2.35.33,34,35).

350
 Appendix I: Expansion of (@B)

Appendix I : The Vector Laplacian in Spherical and Cylindrical Coordinates

I.1 Introduction

This appendix assumes the usual curvilinear coordinates context, Picture B,

. (I.1.1)

In this Section the vector Laplacian is computed in two different ways, each associated with a particular
"tensorization" of its Cartesian form. The second method, though less pleasant than the first, gives insight
into why the vector Laplacian always includes the scalar Laplacian of the field components. It is in this
inclusive form that results are usually stated in the literature.
In passing, it should be noted that the vector Laplacian is not just an idle mathematical curiosity. It
shows up for example in the wave equations for electric and magnetic fields in a vacuum,

(∇2 + k2)E(x) = 0 (∇2 + k2)B(x) = 0 k = ω/c E,B(x,t) = E,B(x)e-iωt (I.1.2)

In continuum mechanics, it appears for example in the Navier/Cauchy equation which describes the small
vector displacement field u in an isotropic elastic solid,

ρo∂t2u = ρ0B + (λ+μ)∇e + μ ∇2u e = div u = dilatation // Lai p 216 (5.6.9) (I.1.3)
or
ρo∂t2u = ρ0B + (λ+μ) (grad div u) + μ ∇2u

Here ρ0 is the unperturbed mass density, B the body force, and λ and μ are the two Lamé constants which
describe an isotropic elastic medium. The vector Laplacian makes another appearance in the better-known
Navier-Stokes equation which describes the vector velocity field v in an incompressible Newtonian fluid,

ρ [ ∂tv + (∇v)v] = ρB - ∇p + μ∇2v // Lai p 361 (6.7.6) (I.1.4)

where ρ is the mass density and p is pressure.

As in previous Appendices, for orthogonal curvilinear coordinates and based on (E.9.13) we can express
the above equations in primed scripted ^
e n-expanded coefficients,

(∇2B)' + k2B' = 0 // vector equation (Helmholtz)

(∇2B)'i + k2B'i = 0 // components (I.1.5)

ρo∂t2u' = ρ0B' + (λ+μ) (grad div u)' + μ (∇2u)' // vector equation (Navier/Cauchy)
ρo∂t2u'i = ρ0B'i + (λ+μ) (grad div u)'i + μ (∇2u)'i // components (I.1.6)

351
 Appendix I: Expansion of (@B)

ρ [ ∂tv' + (∇v)'v'] = ρB' - (∇p)' + μ(∇2v)' // vector equation (Navier-Stokes)

ρ [ ∂tv'i + (∇v)'ijv'j] = ρB'i - (∇p)'i + μ(∇2v)'i // components (I.1.7)

^
In our non-script (e)notation the objects appearing in the above equations may be written

^
u'i = [u(e)]i = h'iu'i // similarly for other vectors

^
(∇p)'i = [(∇p)(e)]i = h'i ∂'ip'(x') where p'(x') = p(x) = scalar field, see (10.1.13)

(grad div u)'i = h'n ∂'n{ (1/ g' ) ∂'i ( g' u 'i/h'i)} // see G in (13.1.17)

^ ^
(∇2B)'i = [(∇2B)(e)]i = [(@B)(e)]i = h'i [(@B)(e)]i = h'i (@B)'i = (@B)'i

= treated in this Appendix. See (I.2.10) for (@B)'i . (I.1.8)

As noted in (E.8.5), the primes on scripted variables correspond to the primes on the names of the
curvilinear coordinates which are x'i for Picture B. In Picture M&S of (14.1.1) the primes on scripted
variables go away, but here we are using Picture B.

I.2 Method 1 : a review

In (13.3.4) it is shown that, in Cartesian coordinates,

∇2(Bn) = [ grad(div B) – curl(curl B) ]n ∇2 = ∂j∂j

or
∇•∇ (Bn) = [∇(∇• B) – ∇ x (∇ x B) ]n . (I.2.1)

When all components are considered in a single equation, one could write

∇2(B) = grad(div B) – curl(curl B)

or
∇•∇ (B) = ∇(∇• B) – ∇ x (∇ x B) (I.2.2)

and this is frequently done (see examples cited in the previous Section). Since ∇2(B) ≠ ∂j∂j B in
curvilinear coordinates, it seems notionally safer in our current context to use a different symbol for the
vector Laplacian operator, and following Moon and Spencer we use @ so the second last equation above
then says:

@B = grad(div B) – curl (curl B) . (I.2.3)

Since [@B]n agrees with ∇2(Bn) in Cartesian coordinates, and since we know how to write div, grad and
curl in curvilinear coordinates (Chapters 9,10,12 or Chapter 15), the right hand side of the above equation

352
 Appendix I: Expansion of (@B)

provides our "first method" of writing @B in curvilinear coordinates. In (15.7.8) it is shown that the
proper "tensorization" of the above equation is given by

(@B)n = (Bj;j);n – g-1/2εnab(g-1/2εbdeBe;d);a @B = (@B)n un (I.2.4)

and therefore, in x'-space (x' are the curvilinear coordinates of interest),

(@B)'n = (B'j;j);n – g'-1/2ε'nab(g'-1/2ε'bdeB'e;d);a @B = (@B)'n en (I.2.5)

where (B'j;j) = [div B]' = [div B] (a scalar). The object @B is a normal vector (weight 0), assuming B is
a normal vector. Doing various simplifying steps, we then arrive at the following expression (15.7.18)
which lends itself to calculation:

(@B)'n = ∂'n{(1/ g' ) ∂'i( g' B'i)} @B = (@B)'n en

– (1/ g' ) ε'ncd ε'eab ∂'c { (1/ g' ) g'de(∂'a[g'bfB'f]) } (I.2.6)

where ε'ncd is the usual permutation tensor (ε'ncd = εncd). In this notation, B'i is an official x'-space
contravariant vector component.
We are often interested in working with vectors which are expanded on unit vector versions of the
tangent base vectors. In such a unit vector expansion of a vector, a script font has been used for the
components. One has,

B'n = B'n/h'n en = h'n ^

en B = B'nen = B'n^
en h'n = scale factor (I.2.7)

which leads to this rewrite of (I.2.6),

(@B)'n = ∂'n{(1/ g' ) ∂'i( g' B'i/h'i)} @B = (@B)'n en (I.2.8)

ncd eab f
– (1/ g' ) ε' ε' ∂'c { (1/ g' ) g'de(∂'a[g'bfB' /h'f]) }

or, as we will use it with unit vectors,

(@B)'n = h'n ∂'n{(1/ g' ) ∂'i( g' B'i/h'i)} @B = (@B)'n ^

en (I.2.9)
ncd eab f
– h'n (1/ g' ) ε' ε' ∂'c { (1/ g' ) g'de(∂'a[g'bfB' /h'f]) } .

Specializing to orthogonal coordinates yields the form we shall use below for computation in Maple,
T1
(@B)' = (1/h'n) ∂'n{(1/ g' ) ∂'i( g' B'i/h'i)}
n
@B = (@B)'n ^ en (I.2.10)
– (h'n/ g' ) ε'ncd ε'dab ∂'c { (1/ g' ) h'd2(∂'a[h'bB'b]) } .
T2
The second term could be written as two terms by reducing the product ε'ncd ε'dab into δδ - δδ in the usual
manner, but Maple is happy to just "do it" as stated. And for non-orthogonal coordinates, the reduction of
ε'ncd ε'eab is much less friendly (see (D.10.18)) and then one would want to use the εε product as is.

353
 Appendix I: Expansion of (@B)

I.3 Method 1 for spherical coordinates: Maple speaks

In this Section, the following notation is used:

B'1 = Br B'2 = Bθ B'3 = Bφ

B = B'n^
e n = Br^
e r + Bθ^
e θ + Bφ^ r + Bθ^
e φ = Br^ ^ .
θ + Bφφ (I.3.1)

Maple begins in the same manner as shown earlier in (G.6.1), the idea being that this code could be
modified for any coordinate system,

354
 Appendix I: Expansion of (@B)

(I.3.2)

The last object is g' , where hp[n] = h'n . The next chunk of code computes the scalar Laplacian of an
unspecified function f, and this expression will be used below in parsing the vector Laplacian results. The
Laplacian expression comes from (11.6) with g'nm = (1/h'n2) δn,m :

(I.3.3)

The subs command used here (and more intensely below) removes the arguments of the function f for
purely cosmetic reasons (see our Maple User's Guide for details, operands section). The arguments are
added in the first place to prevent Maple from thinking the unspecified function f is a constant.
Next comes a low-budget implementation of the permutation tensor eps(a,b,c) = εabc,

355
 Appendix I: Expansion of (@B)

(I.3.4)

The two terms of (@B)'n in (I.2.10) are then duly entered, along with their sum "starB" ,

(I.3.5)

In the following code, we use this notation, similar to (I.3.1) for vector B,

q1_ = (@B)'1 = (@B)r

q2_ = (@B)'2 = (@B)θ
q3_ = (@B)'3 = (@B)φ (I.3.6)

and in this notation Maple computes the components of the vector Laplacian :

356
 Appendix I: Expansion of (@B)

(I.3.7)

See comment below (I.3.3) about the Maple subs commands (purely cosmetic).

I.4 Method 1 for spherical coordinates: putting results in traditional form

It turns out that in each component of the vector Laplacian stated above, 5 of the 9 terms can be
represented as if they were the scalar Laplacian acting on the component in question. Here is how it
works, where we now use the well-known spherical coordinates Laplacian (which Maple has computed
using (11.12) with 1,2,3 = r,θ.φ and h1 = 1, h2 = r, h3 = rsinθ as shown above in (I.3.1) ),

∇2f = (1/r2)∂r(r2∂rf) + (1/r2sinθ)∂θ(sinθ∂θf) + (1/r2sin2θ)∂φ2f =

1+2 3+4 5

1 2 3 4 5 (I.4.1)

357
 Appendix I: Expansion of (@B)

1 2 5 3 4
(I.4.2)
Therefore

(@B)r = ∇2(Br) - (2/r2)Br - (2/r2)cot(θ)Bθ - (2/r2)∂θBθ -(2/r2sinθ) ∂φBφ

= ∇2(Br) – (2/r2) [ Br + cotθ Bθ + ∂θBθ + cscθ ∂φBφ ] (I.4.3)

3 4 5 1 2
(I.4.4)
Therefore

(@B)θ = ∇2(Bθ) - (1/r2) [ - 2 ∂θBr + cot2θ Bθ + Bθ + 2 cotθcscθ∂φBφ ]

= ∇2(Bθ) - (1/r2) [csc2θ Bθ - 2 ∂θBr + 2 cotθcscθ∂φBφ ] (I.4.5)

5 3 4 1 2
(I.4.6)

(@B)φ = ∇2(Bφ) - (1/r2) [csc2θBφ - 2cscθ∂φBr - 2cotθcscθ∂φBθ ] (I.4.7)

Each component is seen to have nine terms.

In this manner, we end up with the components of the vector Laplacian expressed in the traditional
manner,

(@B)r = ∇2(Br) – (2/r2) [ Br + cotθ Bθ + ∂θBθ + cscθ ∂φBφ ]

(@B)θ = ∇2(Bθ) – (1/r2) [csc2θ Bθ – 2 ∂θBr + 2cotθcscθ∂φBφ ]
2 2 2
(@B)φ = ∇ (Bφ) – (1/r ) [csc θ Bφ – 2cscθ∂φBr – 2cotθcscθ∂φBθ ]

358
 Appendix I: Expansion of (@B)

where ∇2f = (1/r2)∂r(r2∂rf) + (1/r2sinθ)∂θ(sinθ∂θf) + (1/r2sin2θ)∂φ2f

= (2/r)∂rf + ∂r2f + (cotθ/r2)∂θf + (1/r2) ∂θ2f + (1/r2sin2θ)∂φ2f (I.4.8)

The curious reader might wonder why, in each case, five of the nine terms of the vector Laplacian
components can be represented by the scalar Laplacian acting on the component. This question is
answered in the following Section.

I.5 Method 2, Part A

In the Method 1, described in Section I.2 above, we used this tensorization of the vector Laplacian,

(@B)'n = (B'j;j);n – g'-1/2ε'nab(g'-1/2g'bcε'cdeB'e;d);a . (I.2.5) (I.5.1)

An alternative tensorization is stated in (15.8.2),

(@B)'n = B'n;j;j . (I.5.2)

These two tensors must be the same since the tensorization of a Cartesian form equation is unique and this
fact is verified in Section 7.8. Our Method 2 is to use this B'n;j;j tensorization to compute once again the
components of the vector Laplacian.

Recall covariant derivative example (F.9.9),

Bab;α ≡ ∂α Bab + Γaαn Bnb + Γbαn Ban (F.9.9)

Since Ba;b is a rank-2 tensor one can apply the above example to Ba;b in place of Bab to get a result of
identical form. The second line below is the desired index contraction (I.5.2) with a→n and b=α → j :

Ba;b;α ≡ ∂α Ba;b + Γaαk Bk;b + Γbαk Ba;k (I.5.3)

Bn;j;j ≡ ∂j Bn;j + Γnjk Bk;j + Γjjk Bn;k // Γ'jjk = (1/ g ) ∂k( g ) as in (F.4.2) .

Next, recall covariant derivative example (F.9.5), with a whole sequence of index shuffles following

Ba;α = ∂α Ba + gαβΓaβn Bn // do β→b and n→s (F.9.5)

Ba;α = ∂α Ba + gαbΓabs Bs // do n→ a and α→j

Bn;j = ∂j Bn + gjbΓnbs Bs // do n→k
Bk;j = ∂j Bk + gjbΓkbs Bs // do k→n and j→k
Bn;k = ∂k Bn + gkbΓnbs Bs (I.5.4)

Insert the last two of these expressions into the second line of (I.5.3) to get,

Bn;j;j = ∂j[∂j Bn + gjbΓnbs Bs] + Γnjk[∂j Bk + gjbΓkbs Bs] + Γjjk[∂k Bn + gkbΓnbs Bs] . (I.5.5)

359
 Appendix I: Expansion of (@B)

Combine the second last term with the first,

Bn;j;j = [∂j ∂j Bn + Γjjk(∂kBn )]

+ ∂j(gjbΓnbs Bs) + Γnjk[∂j Bk + gjbΓkbs Bs] + Γjjk[gkbΓnbs Bs] . (I.5.6)

Since no terms have been dropped, we are still "covariant" and in x'-space everything gets primed,

B'n;j;j = [∂'j ∂'j B'n + Γ 'jjk(∂'kB'n )]

+ ∂'j(g'jbΓ 'nbs B's) + Γ 'njk[∂'j B'k + g'jbΓ 'kbs B's] + Γ 'jjk[g'kbΓ 'nbs B's] (I.5.7)

Using Γ'jjk = (1/ g' ) ∂'k( g' ) from (F.4.2), the first two terms of (I.5.7) can be written this way,

[∂'j ∂'j B'n + Γ 'jjk(∂'kB'n )] = [∂'j ∂'j B'n + (1/ g' ) ∂'k( g' ) (∂'kB'n )] = lap(B'n) (I.5.8)

and we recognize this as being lap (B'n) based on (15.5.6),

lap f = ∂'n∂'nf ' + (1/ g' ) ∂'n( g' ) (∂'nf ') = [1/ g' ] ∂'n [ g' (∂'nf ')] . (15.5.6) (I.5.9)

Therefore (I.5.7) can be written as,

B'n;j;j = lap(B'n) + ∂'j(g'jbΓ 'nbs B's) + Γ 'njk[∂'j B'k + g'jbΓ 'kbs B's] + Γ 'jjk[g'kbΓ 'nbs B's]

= lap(B'n) + Extra Terms (I.5.10)

So we see why the scalar Laplacian of B'n appears in (@B)n. Basically lap(B'n) is the B'n,j,j term within
B'n;j;j , the only term that survives if Γ = 0.

I.6 Method 2, Part B

Unfortunately, we don't want to see lap(B'n) in (I.5.10), we want to see lap(B'n)! Consider then this
rewrite of (I.5.9), using (I.2.7) that B'n = B'n/h'n,

lap (B'n) = lap (B'n/h'n) = [∂'j ∂'j (B''n/h'n) + (1/ g' ) ∂'k( g' ) (∂'k (B''n/h'n) )] . (I.6.1)

One computes the pieces as follows:

∂'j ∂'j (B'n/h'n) = ∂'j ∂'j (B'nh'-1n) = ∂'j [(∂'jB'n) h'-1n + B'n(∂'j h'-1n) ]
= (∂'j∂'jB'n)h'-1n + (∂'jB'n) (∂'j h'-1n) + (∂'j B'n)(∂'j h'-1n) + B'n (∂'j ∂'j h'-1n)
= (∂'j∂'jB'n) h'-1n + 2(∂'jB'n) (∂'j h'-1n) + B'n (∂'j ∂'j h'-1n) (I.6.2)

∂'k (B'n/h'n) = [(∂'kB'n) h'-1n + B'n(∂'k h'-1n) ] (I.6.3)

so that (I.6.1) becomes,

360
 Appendix I: Expansion of (@B)

lap (B'n) = (∂'j∂'jB'n) h'-1n + 2(∂'jB'n) (∂'j h'-1n) + B'n (∂'j ∂'j h'-1n)
+ (1/ g' ) ∂'k( g' )[(∂'kB'n) h'-1n + B'n(∂'k h'-1n) ]

= h'-1n {(∂'j∂'jB'n) + (1/ g' ) ∂'k( g' )(∂'kB'n) }

+ 2(∂'jB'n) (∂'j h'-1n) + B'n (∂'j ∂'j h'-1n) + (1/ g' ) ∂'k( g' )B'n(∂'k h'-1n)

= h'-1n lap (B'n) // (I.5.9) for {..}

j n -1
+ 2(∂' B' ) (∂'j h' n) + B'n (∂'j ∂'j h'-1n) + (1/ g' ) ∂'k( g' )B'n(∂'k h'-1n)

= h'-1n lap (B'n) + Other Terms . (I.6.4)

The conclusion so far from (I.5.2) and (I.5.10) and (I.6.4),

(@B)'n = B'n;j;j = lap(B'n) + Extra Terms (I.5.10)

= [ h'-1n lap(B'n) + Other Terms] + Extra Terms . (I.6.4) (I.6.5)

As before, our interest is with the expansion @B = (@B)'n ^ e n. Since (@B) is a vector like any other
n n
vector, we write (@B)' = (@B)' /h'n as in (I.2.7) so that (I.6.5) becomes,

(@B)'n = lap(B'n) + h'n [Other Terms + Extra Terms] . (I.6.6)

This result applies to any x'-space curvilinear coordinate system, orthogonal or otherwise. Thus we have
demonstrated why it is that lap(B'n) always appears as part of the vector Laplacian component (@B)'n.

We copy the Extras and Other Terms from (I.5.10) and (I.6.4),

Extra Terms = ∂'j(g'jbΓ 'nbs B's) + Γ 'njk[∂'j B'k + g'jbΓ 'kbs B's] + Γ 'jjk[g'kbΓ 'nbs B's] (I.6.7)

Other Terms = 2(∂'jB'n) (∂'j h-1n) + B'n (∂'j ∂'j h-1n) + (1/ g' ) ∂'k( g' )B'n(∂'k h-1n) . (I.6.8)

Rather than attempt algebraic simplification of the above terms, we will just throw them into Maple as is.
Replacing B's = B's/h's in Extra Terms and "lowering" the differential operators appropriately, one gets

Extra Terms =
∂'j(g'jb Γ 'nbs B's/h's) + Γ 'njkg'js∂'s(B'k/h'k) + Γ 'njkg'jb Γ 'kbs B's/h's + Γ 'jjkg'kb Γ 'nbs B's/h's
ET1 ET2 ET3 ET4 (I.6.9)
Other Terms =
2(g'js ∂'sB'n) (∂'jh'n-1) + B'n (∂'j ∂'jh'n-1) + (1/ g' ) ∂'k( g' )B'n(∂'kh'n-1) .
OT1 OT2 OT3 (I.6.10)

361
 Appendix I: Expansion of (@B)

I.7 Method 2 for spherical coordinates: Maple speaks again

In Method 1, there was no need in the Maple calculation for the affine connection object of (F.4.1),

Γ 'dab = (1/2) g'dc [ ∂'ag'bc + ∂'bg'ca – ∂'cg'ab] (I.7.1)

which in orthogonal coordinates simplifies to

Γ 'dab = (1/2) h'd-2 [δdb ∂'a(h'b2) + δda∂'b(h'a2) – δab ∂'d(h'a2)] . (I.7.2)

This last form shows that Γ 'dab = 0 unless two indices match, in which case it might not vanish. For
coordinate systems like spherical and cylindrical coordinates, Γ 'dab is "sparsely populated", and this
explains why won't be swamped by all those Extra Terms shown above.
For spherical coordinates, only 9 of the 27 elements of the object Γdab are non-zero :

Γ '122 = -r Γ '212 = Γ '221 = 1/r // notation: Γ '122 = Γrθθ

Γ '133 = -r sin2θ Γ '313 = Γ '331 = 1/r
Γ '233 = -cosθsinθ Γ '323 = Γ '332 = cotθ // 1,2,3 = r,θ,φ = radius, polar, azimuthal

⎡0 0 0 ⎤ ⎡ 0 1/r 0 ⎤ ⎡ 0 0 1/r ⎤
Γ r = ⎢ 0 -r 0 ⎥ Γθ = ⎢ 1/r 0 0 ⎥ Γφ = ⎢ 0 0 cotθ⎥ (I.7.3)
⎢ 0 0 -rsin2θ ⎥ ⎣ 0 0 -sinθcosθ ⎦ ⎣ 1/r cotθ 0 ⎦
⎣ ⎦

Note that each matrix is symmetric since Γabc = Γacb.

To maintain generality, however, we let Maple compute Γdab = G(d,a,b) from (I.7.1), so

Γ'dab = (1/2) g'dc [ ∂'ag'bc + ∂'bg'ca – ∂'cg'ab] (I.7.4)

The Extra Terms are then entered, as shown above : (I.7.5)

362
 Appendix I: Expansion of (@B)

Extra Terms =
∂'j(g'jb Γ 'nbs B's/h's) + Γ 'njkg'js∂'s(B'k/h'k) + Γ 'njkg'jb Γ 'kbs B's/h's + Γ 'jjkg'kb Γ 'nbs B's/h's
ET1 ET2 ET3 ET4

(I.7.6)
And then the Other Terms are entered as well,

Other Terms =
2(g'js ∂'sB'n) (∂'jh'n-1) + B'n (∂'j ∂'jh'n-1) + (1/ g' ) ∂'k( g' )B'n(∂'kh'n-1) (I.7.7)
OT1 OT2 OT3

(I.7.8)

363
 Appendix I: Expansion of (@B)

Maple now computes the non- lap(B'n) terms shown in (I.6.6), called starBx(n) in the code,

(@B)'n = lap(B'n) + h'n [Other Terms + Extra Terms] . (I.6.6)

The final results are then generated :

(I.7.9)
Recall that results of Method 1 were,

(@B)r = ∇2(Br) – (2/r2) [ Br + cotθ Bθ + ∂θBθ + cscθ ∂φBφ ]

(@B)θ = ∇2(Bθ) – (1/r2) [csc2θ Bθ – 2 ∂θBr + 2cotθcscθ∂φBφ ]
2 2 2
(@B)φ = ∇ (Bφ) – (1/r ) [csc θ Bφ – 2cscθ∂φBr – 2cotθcscθ∂φBθ ] (I.4.8)

One then sees that the non-scalar-Laplacian terms just computed as (I.7.9) by Method 2 exactly match the
non-∇2 terms in (I.4.8) as found by Method 1.

I.8 Results for Cylindrical Coordinates from both methods

The Maple program was easily modified for this system. Here are the results:

B'1 = Br B'2 = Bφ B'3 = Bz

B = B'n^
e n = Br^
e r + Bφ^
e φ + Bφ^
e z = Br^ ^ + Bφ^
r + Bθφ z . (I.8.1)

The scalar Laplacian of an unspecified function f :

∇2f = (1/r)∂r(rf) + (1/r2)∂φ2f + ∂z2f =

1+2 3 4

364
 Appendix I: Expansion of (@B)

(I.8.2)
1 2 3 4

The components of the vector Laplacian are found by Method 1 to be

(I.8.3)
1 2 4 3
Therefore,

(@B)r = ∇2(Br) - Br/r2 - 2∂φBφ/r2 . (I.8.4)

(I.8.5)
3 4 1 2
Therefore,

(@B)φ = ∇2(Bφ) - Bφ/r2 + 2∂φBr/r2 . (I.8.6)

(I.8.7)
4 3 1 2

Therefore,

(@B)z = ∇2(Bz) (I.8.8)

365
 Appendix I: Expansion of (@B)

as befits a component which is Cartesian. In this manner, Method 1 produces the components of the
vector Laplacian expressed in the traditional manner,

(@B)r = ∇2(Br) – (1/r2)[Br + 2∂φBφ]

(@B)φ = ∇2(Bφ) – (1/r2)[Bφ – 2∂φBr]
(@B)z = ∇2(Bz)

where ∇2f = (1/r)∂r(rf) + (1/r2)∂φ2f + ∂z2f = ∂2rf +(1/r)∂rf + (1/r2)∂φ2f + ∂z2f (I.8.9)

Method 2 produces the following results for the terms which are added to the scalar Laplacian,

(I.8.10)
2
and these are seen to agree with non-∇ terms in (I.8.9),

(@B)r = ∇2(Br) – (1/r2)[Br + 2∂φBφ]

(@B)φ = ∇2(Bφ) – (1/r2)[Bφ – 2∂φBr]
(@B)z = ∇2(Bz) (I.8.9)

In cylindrical coordinates the Γ object is even sparser than in spherical coordinates. One has

Γ 'dab = (1/2) h'd-2 [δdb ∂'a(h'b2) + δda∂'b(h'a2) – δab ∂'d(h'a2)] .

Γ '122 = -r // notation: Γ '122 = Γrφφ

Γ '212 = 1/r
Γ '221 = 1/r // 1,2,3 = r,φ,z
r φ z
⎡0 0 0⎤ ⎡ 0 1/r 0⎤ r ⎡0 0 0⎤
Γ = ⎢ 0 -r 0 ⎥
r
Γ = ⎢ 1/r
φ
0 0⎥ φ Γ = ⎢ 0 0 0⎥
z
(I.8.11)
⎣0 0 0⎦ ⎣ 0 0 0⎦ z ⎣0 0 0⎦

366
 Appendix I: Expansion of (@B)

so only 3 of 27 components are non-vanishing.

Hand Calculation of Γ' for cylindrical coordinates

From a simple polar coordinates picture one knows that

d^ ^
r = dφ φ ^ = -dφ ^
dφ r. (I.8.12)

Then from (3.4.2) with θ → φ, we have this situation for cylindrical coordinates,

er = ^
er = ^
r ⇒ der = d^
r = dφ φ^ = dφ (eφ/r) = (1/r)dφ eφ
^
eφ = r φ ⇒ ^ ) = r dφ
deφ = d(r φ ^ + dr φ
^ = r ( -dφ ^ ^ = -rdφ er + (dr/r)eφ
r ) + dr φ
ez = ^
z ⇒ dez = 0 . (I.8.13)

Therefore only 3 of 9 partial derivatives are non-zero,

∂rer = 0 ∂reφ = (1/r)eφ ∂zer = 0

∂φer = (1/r)eφ ∂φeφ = -rer ∂zeφ = 0
∂zer = 0 ∂zeφ = 0 ∂zez = 0 . (I.8.14)

Since we are in the Picture A context of Fig (3.4.3), we use (F.1.11) which says

Γ'cab = ec • (∂'aeb) . (F.1.11)

Then

Γ'cab = g'cd ed • (∂'aeb) = h'c-2 ec • (∂'aeb) . // no sum on c (I.8.15)

For our situation with x' = (r,φ,z), we find that for Γ'cab to not vanish, one must have ab = φr, rφ and φφ
since all other (∂'jeb) = 0 in (I.8.14). In each of these cases only one Γ component survives due the
orthogonality of the base vectors :

Γ'φφφ = hφ-2 eφ • (∂φeφ) = r-2 eφ • [ -rer] = 0

Γ'zφφ = hz-2 ez • (∂φeφ) = 1 * ez • [ -rer] = 0

Γ'rφφ = hr-2 er • (∂φeφ) = 1 * er • [ -rer] = -r ^

r •^
r =-r

^•φ
Γ'φrφ = hφ-2 eφ • (∂reφ) = r-2 eφ • [(1/r)eφ] = r-2 r φ ^ = (1/r)

^•φ
Γ'φφr = hφ-2 eφ • (∂φer) = r-2 eφ • [ (1/r)eφ] = r-2 r φ ^ = (1/r) (I.8.16)

These results agree with (I.8.11).

367
 Appendix J: Expansion of (∇T)

Appendix J: Expansion of (∇T) in curvilinear coordinates (T = rank-2 tensor)

J.1 Total time derivative as prototype equation

This appendix assumes the usual curvilinear coordinates context, Picture B :

(J.1.1)

The total time derivative of a contravariant rank-2 tensor field Tij(x,t) can be written as

dTij(x,t)/dt = ∂Tij/∂t + (∂Tij/∂xk) (∂xk/dt) = ∂Tij/∂t + (∂Tij/∂xk) vk

or
dtTij = ∂tTij + (∂kTij) vk . // dt ≡ d/dt, ∂t ≡ ∂/∂t, ∂k ≡ ∂/∂xk (J.1.2)

We take this as a useful prototype equation for two reasons. First, it contains our object of interest, which
is the gradient of a rank-2 tensor, ∂kTij. Second, this equation plays a role in the continuum mechanics of
non-Newtonian fluids as discussed in Section J.6 below.
One can define, in Cartesian coordinates, a (∇T) object :

(∇T)ijk ≡ ∂kTij (J.1.3)

so that, from (J.1.2),

dtTij = ∂tTij + (∇T)ijk vk . (J.1.4)

Comment: Our convention has been to bold vectors and not to bold other tensors. In line with this idea,
we shall write ∇T where the grad is bolded and the T is not bolded.

In order to express the above equation in curvilinear coordinates, it must be "tensorized" in the sense of
Section 15.2 so that the equation is covariant. Thus, (∇T)ijk must be regarded as components of a
(mixed) rank-3 tensor which, in Cartesian coordinates, are equal to ∂kTij. Since vk are the components of
a tensorial vector, (∇T)ijkvk transforms as a rank-2 tensor, and then all terms in the above equation are
rank-2 tensors and the equation is then covariant and therefore appears this way in x'-space,

dtT'ij = ∂tT'ij + (∇T)'ijk v'k . (J.1.5)

In our usual formalism, x'-space is the space of some generic curvilinear coordinates x'n (not necessarily
orthogonal) and then (J.1.5) tells us the form taken by the time derivative equation in curvilinear
coordinates, and it remains only to compute the objects (∇T')ijk.

368
 Appendix J: Expansion of (∇T)

For convenience, we can lower the tensorial ij indices on the above tensor equations to get

dtTij = ∂tTij + (∇T)ijk vk (∇T)ijk ≡ ∂kTij // Cartesian coordinates

dtT'ij = ∂tT'ij + (∇T)'ijk v'k // curvilinear coordinates (J.1.6)

and then we can deal with the pure covariant tensor components (∇T)'ijk .

As in previous appendices, we shall be interested in (J.1.6) stated in ^

e n-expanded coefficients according
to the covariance idea of (E.9.13),

dtT'ij = ∂tT'ij + (∇T)'ijk v'k (J.1.7)

^
where the components (∇T)'ijk = (∇T)'ijk = [(∇T)(e)]ijk are computed in (J.7.8) below and evaluated
for specific orthogonal coordinate systems in subsequent equations.

J.2 Computation of components (∇T)'ijk

The covariant derivative is discussed in Sections F.7- F.9. We shall define a true tensor object (∇T)abc as
Tab;α which is defined in (F.9.6),

(∇T)abc ≡ Tab;c ≡ Tab,c – ΓnacTnb – ΓnbcTan = Tab,c = ∂cTab // x-space

(∇T)'abc ≡ T'ab;c ≡ T'ab,c – Γ'nacT'nb – Γ'nbcT'an . // x'-space (J.2.1)

Recall that Tab,c is a shorthand for ∂cTab and that Γcab is the affine connection which tells how the basis
vectors change as one moves around in space. In Cartesian x-space (first line above, see also Picture C1 in
Fig (F.1.1)) the basis vectors are the fixed un which don't change, so Γcab ≡ 0 as in (F.4.16). In
curvilinear x'-space, Γ'cab ≠ 0. Since Tab;c transforms as a true rank-3 tensor, its defining equation is
"covariant" (Section 7.15) so that in x'-space the equation has exactly the same form but everything is
primed.
The above two lines characterize the process of "tensorization": we find a true "tensorial tensor"
Tab;c which agrees with (∇T)abc = ∂cTab in Cartesian space. The tensorized version of Tab,c is unique
(as shown in Section 15.2), and it is Tab;c . Since this tensor is given by T'ab;c in x'-space, we use the
second line above to compute the components of the tensor (∇T) object in x'-space, which is to say, in
curvilinear coordinates.
It is a simple matter to have Maple compute Γ'cab for any curvilinear coordinate system, and then the
second line in (J.2.1) reports out the components (∇T)'abc,

(∇T)'abc = ∂'cT'ab – Γ'nacT'nb – Γ'nbcT'an

Γ 'dab = (1/2) g'dc [ ∂'ag'bc + ∂'bg'ca – ∂'cg'ab] // (F.4.1) (J.2.2)

369
 Appendix J: Expansion of (∇T)

where g'ab is the metric tensor in x'-space. Then we know from (J.1.6) how to write the time derivative
equation in any curvilinear coordinates,

dtT'ij = ∂tT'ij + (∇T)'ijk v'k

or
dtT'ij = ∂tT'ij + (∇T)'ijk v'k (∇T)'ijk = g'ii'g'jj'(∇T)'i'j'k . (J.2.3)

In (I.3.2) we show Maple code to compute the covariant metric tensor g'ij from the curvilinear
coordinates' defining equations (such as x = rsinθcosφ for sphericals). Then (I.7.5) shows the extra code
for computing g'ij and Γ 'dab . The reader can then add a few extra lines to have Maple compute the
desired (∇T)'abc using the equations (J.2.2) above. Later in this Appendix we shall use Maple to compute
certain related quantities which we can then verify against a known source.

J.3 Tensor expansions of ∇T on the un and en base vectors

It is useful at this point to write out the tensor expansions for the rank-3 tensor (∇T) to show exactly
where the components (∇T)'abc appear. As shown in (E.2.11) and (E.2.14) one can expand the rank-3
tensor (∇T) in various ways. The first line below shows expansions on x-space basis vectors, while the
second line shows expansion on the reciprocal and tangent base vectors (which also exist in x-space) :

∇T = Σijk Tij;k ui⊗uj⊗uk = Σijk [(∇T)(u)]ijk ui⊗uj⊗uk = Σijk [(∇T)(u)]ijk ui⊗uj⊗uk

∇T = Σijk T'ij;k ei⊗ej⊗ek = Σijk [(∇T)(e)]ijk ei⊗ej⊗ek = Σijk [(∇T)(e)]ijk ei⊗ej⊗ek .
(J.3.1)
As usual, up and down index positions on T are the same for the first line in Cartesian space, but are
significant on the second line (for non-orthogonal coordinates). The superscripts on the (∇T)(u) and
(∇T)(e) components indicate which basis vectors are being expanded upon. One normally just writes
(∇T)(u) = (∇T), and (∇T)(e) = (∇T)' where the prime indicates the curvilinear x'-space. So, we now have
three different notations for the expansion components:

Tij;k = [(∇T)(u)]ijk = (∇T)ijk

T'ij;k = [(∇T)(e)]ijk = (∇T)'ijk . (J.3.2)

Just to fill things out, here are the corresponding expansions for rank-2 tensor T, again from (E.2.11) and
(E.2.14),

T = Σij Tij ui⊗uj = Σij [T(u)]ij ui⊗uj = Σij [T(u)]ij ui⊗uj Tij = [T(u)]ij
T = Σij T'ij ei⊗ej = Σij [T(e)]ij ei⊗ej = Σij [T(e)]ij ei⊗ej T'ij = [T(e)]ij (J.3.3)

and then for a rank-1 tensor, as in (7.13.10),

v = Σi vi ui = Σi [v(u)]i ui = Σi [v(u)]i ui = Σi viui vi = [v(u)]i

v = Σi v'i ei = Σi [v(e)]i ei = Σi [v(e)]i ei = Σi v'iei v'i = [v(e)]i . (J.3.4)

370
 Appendix J: Expansion of (∇T)

J.4 Tensor expansions of ∇T on the ^

e n base vectors

In practical applications, it is sometimes useful to deal with components of tensors which are expanded on
the unit versions of the tangent base vectors, ^ e n ≡ en/ |en| = en/h'n . On the one hand, this introduces
major complications (see below) since such components are non-covariant (neither contravariant nor
covariant nor mixed). On the other hand, since the unit vectors are all dimensionless, all tensor
components have the same dimensions, which is very useful in any practical engineering work. We shall
refer to such tensor components as "unit-base-vector components".

This subject is addressed below (E.2.14) and in Section E.8. We start with one of the (J.3.1) expansions,

∇T = Σijk [(∇T)(e)]ijk ei⊗ej⊗ek = Σijk (∇T)'ijk ei⊗ej⊗ek (J.4.1)

and process it in this manner ,

= Σijk [(∇T)(e)]ijk (h'i^

e i) ⊗ (h'j^
e j) ⊗ (h'k^
e k)

= Σijk { h'i h'j h'k [(∇T)(e)]ijk } ^

e i⊗^
e j ⊗^
ek

^ ^
= Σijk [(∇T)(e)]ijk ^
e i ⊗^
e j⊗^
e k where [(∇T)(e)]ijk = h'i h'j h'k [(∇T)(e)]ijk . (J.4.2)

If our x'-space coordinate system happens NOT to be orthogonal, then the up/down position of the indices
^
on [(∇T)(e)]ijk has significance. This is so because, as shown near (E.2.11), these indices are lowered
by w'nm = ^ en • ^e m . For an orthogonal system, one has w'nm = ^
en • ^e m = δn,m and then the up/down
^
positions of the indices on [(∇T)(e)]ijk all indicate the same number. One should not confuse this with
the fact that the up/down position of an index on any true x'-space tensor like (∇T)'ijk is always
significant, whether or not a coordinate system is orthogonal, because in general g'ab ≠ δa,b.
Now for convenience later, we make one more definition,

^
(∇T)'ijk ≡ [(∇T)(e)]ijk = h'i h'j h'k [(∇T)(e)]ijk = h'i h'j h'k (∇T)'ijk (J.4.3)

where we follow our convention that the components of unit-base-vector tensors are written in script, see
(E.8.5). (The symbol T is a script T , not a "tau" τ. )
Similarly, we can write the unit-base-vector expansion for a rank-2 tensor T,

^ ^
T = Σij [T(e)]ij ^
e i⊗^
e j where [T(e)]ij = h'i h'j [T(e)]ij (J.4.4)

with the definition

^
T'ij ≡ [T(e)]ij = h'i h'j [T(e)]ij = h'i h'j T'ij . (J.4.5)

Finally, for a vector v,

371
 Appendix J: Expansion of (∇T)

^ ^
v = Σi [v(e)]i ^
e i where [v(e)]i = h'i [v(e)]i (J.4.6)

with the definition

^
v'i ≡ [v(e)]i = h'i [v(e)]i = h'i v'i . (J.4.7)

In terms of the scripted unit-base-vector components, our three expansions are:

∇T = Σijk (∇T)'ijk ^
e i ⊗^
e j⊗^
ek (∇T)'ijk = h'i h'j h'k (∇T)'ijk

T = Σij T'ij ^
e i ⊗^
ej T'ij = h'i h'j T'ij

v = Σi v'i ^
ei v'i = h'i v'i . (J.4.8)

The scripted tensor components have indices which are integers, i = 1,2...N. Once we actually select a
particular curvilinear coordinate system, one can replace the indices with curvilinear coordinate names
and then, since those names indicate that one is talking about x'-space components, and since we just
assume we are dealing with unit-base-vector components, both the script and the prime position can be
dropped. For example, in spherical coordinates with 1,2,3 = r,θ,φ one can write

(∇T)'123 = (∇T)rθφ
T'12 = Trθ T'11 = Trr
v'3 = vφ v'1 = vr (J.4.9)

As noted below (J.4.2), if a curvilinear coordinate system is orthogonal (and this is generally the case, and
is certainly the case for spherical coordinates), then the up/down position does not matter and one writes

(∇T)'123 = (∇T)'123 = (∇T)rθφ

T'12 = Trθ T'11 = Trr
v'3 = vφ v'1 = vr . (J.4.10)

Notice that the scripted forms are always necessary when summations like Σijk are involved, unless one
is willing to write out all the terms in the sum, which is a bit clumsy. The continuum mechanics book of
Lai et al., which we shall refer to below, avoids summations in curvilinear coordinates and thus has no
need for our scripted components.

J.5 Total time derivative equation written in unit-base-vector curvilinear components

Recall from (J.2.3) our prototype time derivative equation of interest in x'-space,

dtT'ij = ∂tT'ij + (∇T)'ijk v'k . (J.5.1)

372
 Appendix J: Expansion of (∇T)

How does one write this equation in terms of unit-base-vector tensor components? We know that for
orthogonal coordinates the answer is (J.1.7) based on covariance. Here we shall answer the question for
general coordinates, and shall show that (J.1.7) "obtains" for orthogonal coordinates.

From (J.4.5) and (J.4.7) we had (no implied sums here)

T'ij = h'i h'j T'ij v'k = h'k v'k (J.5.2)

so (J.5.1) can be processed as follows:

dtT'ij = ∂tT'ij + (∇T)'ijk v'k // (J.5.1)

h'i h'j dtT'ij = h'i h'j ∂tT'ij + h'i h'j (∇T)'ijk h'k-1 h'k v'k // mult thru by h'i h'j

dt(h'ih'j T'ij) = ∂t(h'ih'j T'ij) + h'i h'j h'k-1 (∇T)'ijk (h'k v'k) // h'n = h'n(x'), no t

dt T'ij = ∂t T'ij + [ h'i h'j h'k-1 (∇T)'ijk ] v'k // (J.5.2)

dt T'ij = ∂t T'ij + Q'ijk v'k Q'ijk ≡ h'i h'j h'k-1 (∇T)'ijk . (J.5.3)

Comment: Note that dth'n(x) = ∂th'n(x) = 0 and not dth'n(x) = ∂th'n(x) + (∇h'n) v. The reason is that the
field h'n(x) is not an Eulerian fluid property like A1 in (J.6.9) below, it is a property of space at point x.

Recall that no assumption has been made that the curvilinear coordinates are orthogonal. Section J.2
showed how the (∇T)'ijk can be computed for any curvilinear coordinate system, and thus one can
compute the Q'ijk shown above. Our question is now answered: (J.5.3) shows how one writes the total
time derivative equation in arbitrary curvilinear coordinates.

If we now assume the coordinates are orthogonal, then

(∇T)'ijk = gkk'(∇T)'ijk' = hk2δk,k' (∇T)'ijk' = hk2(∇T)'ijk (J.5.4)

and then

Q'ijk ≡ h'i h'j h'k-1 (∇T)'ijk = h'i h'j h'k-1 hk2(∇T)'ijk

= h'i h'j h'k(∇T)'ijk = (∇T)ijk . // (J.4.3) (J.5.5)

As noted below (J.4.2), for orthogonal coordinates the up/down index position on (∇T)ijk makes no
difference, so for example, (∇T)ijk = (∇T)ijk . Installing Q'ijk = (∇T)ijk into (J.5.3) then gives,

dt T'ij = ∂t T'ij + (∇T)'ijk v'k (∇T)'ijk ≡ h'i h'j h'k (∇T)'ijk . // orthogonal (J.5.6)

373
 Appendix J: Expansion of (∇T)

which is (J.1.7) which was deduced by covariance. Once again, indices go up and down for free on T'ij
and (∇T)ijk, but not on T'ij and (∇T)'ijk . We shall assume from now on that the curvilinear coordinates
are orthogonal.
As a reminder, here is what the above equation says if i = j = 1,

dt T'11 = ∂t T'11 + (∇T)'11k v'k . (J.5.7)

Using the notation scheme just described above in (J.4.9) and (J.4.10), in spherical coordinates one would
write the above equation as,

dt Trr = ∂t Trr + (∇T)rrr vr + (∇T)rrθ vθ + (∇T)rrφ vφ . (J.5.8)

J.6 Shorthand notations and a continuum mechanics application

Consider our original Cartesian time derivative equation (J.1.2),

dtTij = ∂tTij + (∇T)ijk vk . (J.6.1)

One could regard (∇T)ijk as the kth component of a vector (∇T)ij labeled by fixed values i and j, so
that

[(∇T)ij]k = (∇T)ijk . (J.6.2)

Then one can write (J.6.1) as

dtTij = ∂tTij + (∇T)ij • v . (J.6.3)

The next step is to suppress the ij labels, since the equation above is true for any i and j,

dtT = ∂tT + (∇T) • v . (J.6.4)

This is only a shorthand notation, no precision justification is required. We can do the same thing to the
equation written for orthogonal unit-base-vector curvilinear coordinates :

dt T'ij = ∂t T'ij + (∇T)'ijk v'k (J.5.6)

[(∇T)'ij]k = (∇T)'ijk // vector with index k

dt T'ij = ∂t T'ij + (∇T)'ij • v'k // dot notation

dt T' = ∂t T' + (∇T)' • v' // indices stripped (J.6.5)

and then we have

374
 Appendix J: Expansion of (∇T)

dtT = ∂tT + (∇T) • v

dt T' = ∂tT' + (∇T)' • v' . (J.6.6)
In the book of Lai et al., the • is omitted and the shorthand notations are written,

dtT = ∂tT + (∇T) v

dt T' = ∂tT' + (∇T)' v' (J.6.7)

where one imagines that (∇T) and (∇T)' are 3-index operators which act on a 1-index object to generate a
2-index object. Again, it is just a shorthand. The real meaning of these equations is

dtTij = ∂tTij + (∇T)ijk vk

dt T'ij = ∂tT'ij + (∇T)'ijk v'k . (J.6.8)

Example: An application of our total time derivative equation appears in the first line of Lai p 470 ,

DA1/Dt = ∂A1/∂t + (∇A1) v (J.6.9)

where D/Dt is the way a total time derivative is expressed in continuum mechanics (D/Dt = d/dt). It is the
convective or material derivative for Eulerian-picture functions (like A1(x,t)), meaning that the second
term registers a time change in a fluid property at the fixed point x due to "new fluid" with velocity v
passing through that point (a point being a tiny differential volume of fluid). In the notation of (J.6.7), this
would be written

dtA'1 = ∂tA'1 + (∇A1)' v' . (J.6.10)

The detailed meaning is

dt (A'1)ij = ∂t (A'1)ij + (∇A1)'ijk v'k (J.6.11)

and for i = j = 1 this says (spherical coordinates) ,

dt[(A1)rr] = ∂t[(A1)rr] + (∇A1)rrr vr + (∇A1)rrθ vθ + (∇A1)rrφ vφ . (J.6.12)

The tensor A1 is the "first Rivlin-Ericksen tensor" associated with the flow of non-Newtonian fluids
(rheology). The Ai tensors, briefly mentioned in Section K.4, are derivatives of a certain deformation
tensor called Ct, and computation of the Ai is done in the following iterative manner,

Ai+1 = DtAi + Ai(∇v) + (∇v)TAi , // Lai p 468 (8.11.3) (J.6.13)

where (∇v) is the gradient-of-vector object treated in our Appendix G, v being the fluid velocity field. In
particular, A2 = DtA1 + A1(∇v) + (∇v)TA1 which involves our object of interest DtA1 = dtA1. So, in
order to compute A2 in curvilinear coordinates, one needs dt(A'1)ij which involves the (∇A1)'ijk. In
order to use the above equation in practice, one has to know for example that (∇A1)rrθ = [ ∂θ(A1)rr -
(A1)θr - (A1)rθ]/r, a fact that is certainly not immediately obvious (but see the rrθ entry in (J.7.9) below).

375
 Appendix J: Expansion of (∇T)

So, our next task is to compute the (∇T)'ijk which appear in our prototype equation above,

dt T' = ∂t T' + (∇T)' v' // shorthand notation (J.6.7)

dt T'ij = ∂t T'ij + (∇T)'ijk v'k (J.5.6) (J.6.14)

J.7 Maple computation of the (∇T)'ijk components for spherical coordinates

Recall from (J.4.3) that, for arbitrary curvilinear coordinates,

(∇T)'ijk ≡ h'i h'j h'k (∇T)'ijk (J.7.1)

and from (J.2.2),

(∇T)'abc = ∂'cT'ab – Γ'nacT'nb – Γ'nbcT'an (J.7.2)

Γ 'dab = (1/2) g'dc [ ∂'ag'bc + ∂'bg'ca – ∂'cg'ab] . g' = metric tensor for x'-space

But we are now assuming only orthogonal coordinates, so

g'ab = δa,bha2 g'ab = δa,bha-2 (J.7.3)

⇒ (∇T)'ijk = g'ii'g'jj'g'kk'(∇T)'i'j'k' = (h'i h'j h'k)-2 (∇T)'ijk (J.7.4)

and then from (J.7.1),

(∇T)'ijk = (∇T)'ijk = (h'i h'j h'k)-1 (∇T)'ijk . (J.7.5)

Remember that up/down index position does not matter on (∇T)'ijk for orthogonal coordinates.

We want the result expressed in terms of the T'ij and not the T'ij, so from (J.4.8),

T'ij = (h'i h'j) T'ij = (h'i h'j) g'ii' g'jj'T'i'j' = (h'i h'j)-1 T'ij

⇒ T'ij = (h'i h'jT'ij) . (J.7.6)

Then (J.7.2) reads,

(∇T)'abc = ∂'cT'ab – Γ'nacT'nb – Γ'nbcT'an // do abc→ijk rename

(∇T)'ijk = ∂'kT'ij – Γ'nikT'nj – Γ'njkT'in

(∇T)'ijk = ∂'k(h'i h'jT'ij) – Γ'nik(h'n h'jT'nj) – Γ'njk(h'i h'nT'in) // (J.7.6)

376
 Appendix J: Expansion of (∇T)

= h'i h'j(∂'kT'ij) + ∂'k(h'i h'j) T'ij – Γ'nik(h'n h'jT'nj) – Γ'njk(h'i h'nT'in) (J.7.7)

where we break the ∂'k term in two pieces for Maple technical reasons.

So the Maple program will compute from (J.7.5) and (J.7.7),

(∇T)'ijk = (h'i h'j h'k)-1 (∇T)'ijk (J.7.8)

where (∇T)'ijk = h'i h'j(∂'kT'ij) + ∂'k(h'i h'j) T'ij – Γ'nik(h'n h'jT'nj) – Γ'njk(h'i h'nT'in) .
T1 T2 T3 T4

For now we assume spherical coordinates. Maple first computes g** = gcov using (I.3.2) and g** =
gcontra using (I.7.5). Then the affine connection is computed from these metric tensors, just as in (I.7.5),

(J.7.9)

Then the terms of (J.7.8) are entered (T'ij = Te[i,j], h'i = hp[i], Γ 'dab = G(d,a,b), etc.)

(J.7.10)

The terms are then added, multiplied by (h'i h'j h'k)-1 according to (J.7.8), and then displayed. In this
code then, DT(i,j,k) is (∇T)'ijk as computed on the first line. In the triple loop, the complicated resulting
expressions are cleaned up in a series of cosmetic trigonometric filtering steps: DT(i,j,k)→ f → g → h[k]
and it is then the simplified h[k] expressions which are printed out.

377
 Appendix J: Expansion of (∇T)

(J.7.11)
and here are the resulting values for (∇T)'ijk (for example, (∇T)'112 = (∇T)rrθ )

(∇T)'ijk for Spherical Coordinates (J.7.12)

378
 Appendix J: Expansion of (∇T)

The strange " symbols in the above table should be ignored, just a Maple print command artifact.
These results agree with the spherical coordinates table given in Lai p 505.
The divT results of Appendix H can be verified from the above table using

(divT)'i = ∂'jT'ij = (∇T)'ijj = Σj(∇T)'ijj . (J.7.13)

For example,

(divT)r = (∇T)rrr + (∇T)rθθ + (∇T)rφφ

= ∂rTrr + (1/r)[∂θTrθ - Tθθ + Trr] + (1/r)[ cscθ ∂φTrφ - Tφφ + Trr + cotθ Trθ]

= ∂rTrr + (1/r)∂θTrθ - Tθθ/r + (2/r)Trr + (1/rsinθ) ∂φTrφ - Tφφ/r + cotθ Trθ /r (J.7.14)

and we quote from (H.6.3),

. (H.6.3)

379
 Appendix J: Expansion of (∇T)

J.8 Maple computation of the (∇T)'ijk components for cylindrical coordinates

Making four small edits to the spherical coordinates Maple program converts it to a cylindrical
coordinates program. Here are the resulting values for (∇T)'ijk (for example, (∇T)'123 = (∇T)rθz )

(∇T)'ijk for Cylindrical Coordinates (J.8.1)

and these expressions agree with those in the table on page 504 of Lai. The same comment made about
divT at the end of the previous Section applies here as well. The object Γ 'dab has 3 of 27 components non
zero as shown in (I.8.11), only 2 of which are distinct.

380
 Appendix J: Expansion of (∇T)

It should be emphasized that this same simple Maple code can be used to compute the (∇T)'ijk for any
system of orthogonal curvilinear coordinates in any number of dimensions N. For non-orthogonal
coordinates the results would be obtained from these more general equations,

(∇T)'ijk = h'i h'j h'k (∇T)'ijk (J.7.1)

(∇T)'ijk = g'ai g'bj g'ck (∇T)'abc // raise indices

(∇T)'abc = ∂'cT'ab – Γ'nacT'nb – Γ'nbcT'an (J.7.2)

Γ 'dab = (1/2) g'dc [ ∂'ag'bc + ∂'bg'ca – ∂'cg'ab] . g' = metric tensor for x'-space
(J.8.2)

J.9 The Lai Method of computing (∇T)'ijk for orthogonal coordinates

This method appears in Lai pp 501-505. It differs from the method given above in Section J.8 mainly
because it uses a different affine connection, but of course it gives the same results. We shall present Lai's
computation of the components (∇T)'ijk in the context of Picture B,

(J.9.1)

At the very end we will translate the result to Picture C1 of (F.1.1) where x instead of x' are the
curvilinear coordinates.
For Picture B, one has from (F.1.11),

(∂'jen) = Γ'kjn ek ⇒ (den)i = (∂'jen)i dx'j = Γ'kjn (ek)i dx'j (J.9.3)

Γ'kjn = ek • (∂'jen) (J.9.4)

where the en are our usual Chapter 3 tangent base vectors and we regard en = en(x').
^'k in this manner,
It is possible to define a different affine connection Γ jn

(∂'j^ ^'k ^ (d^ ^'k (e

e n)i = Γ jn ^k) dx'
i j
e n) = Γ jn e k ⇒ (J.9.4)

^
Γ'kjn = ^
e k • (∂'j^
e n) . (J.9.5)

^'k describes how the unit vectors ^

This new affine connection Γ e n vary with x'.
jn

Now consider the tensor expansion from (J.4.8), where all objects are functions of x',

381
 Appendix J: Expansion of (∇T)

T = Σij T'ij ^
e i ⊗^
e j = T'ij ^
e i⊗^
ej . (J.9.6)

The object T = T(x') is a vector in a double direct-product space spanned by ^

e i⊗^
e j. Apply ∂'k to get

∂'kT = (∂'kT'ij) ^
e i⊗^
e j + T'ij(∂'k^
e i)⊗^
e j + T'ij ^
e i⊗(∂'k^
e j) . (J.9.7)

Now do index shuffles on (J.9.4),

(∂'j^ ^'k ^
e n) = Γ // do k→s then j → k and n → i
jn e k

(∂' ^
e )=Γ ^'s ^
e // do i→j
k i ki s

(∂'k^ ^'s ^
e j) = Γ kj e s . (J.9.8)

Insert the last two lines into (J.9.7) to get

∂'kT = (∂'kT'ij) ^ e j + T'ij^

e i⊗^ Γ'ski ^
e s ⊗^
e j + T'ij ^ ^'s ^
e i⊗ Γ kj e s

In the second term swap s↔ i and in the third term swap s↔ j (dummy summation indices) to get,

∂'kT = (∂'kT'ij) ^ e j + T'sj^

e i⊗^ Γ'iks ^
e i ⊗^
e j + T'is ^ ^'j ^
e i⊗ Γ ks e j

= [∂'kT'ij + T'sj^
Γ'iks + T'is^
Γ'jks ] ^
e i ⊗^
e j, (J.9.9)

so ∂'kT is another vector in the same double direct-product space. To economize below, we write the
above as

∂'kT = Qijk ^
e i⊗^
ej

^'i + T'is^
where Qijk ≡ [∂'kT'ij + T'sjΓ Γ'jks ] . (J.9.10)
ks

Next, dot both sides of (J.9.10) with the vector ua ⊗ ub where the un are our usual Cartesian space axis-
aligned unit vectors (un)i = (un)i = δi,n . On the left side one gets

(∂'kT) • ua ⊗ ub = (∂'kT)ab = ∂'kTab (J.9.11)

where Tab are the contravariant components of the T tensor in x-space. On the right side one gets

^
e i⊗^
e j • ua ⊗ ub = (^
e i•ua) (^
e j•ub) = (^
e i)a (^
e j)b . (J.9.12)

Thus (J.9.10) becomes

382
 Appendix J: Expansion of (∇T)

∂'kTab = Qijk (^
e i)a (^
e j)b . (J.9.13)

We know from (7.6.8) that ∂c = Rkc ∂'k, so applying Rkc to both sides of (J.9.13) gives

∂cTab = Rkc (^
e i)a (^
e j)b Qijk . (J.9.14)

Since x-space is Cartesian, this is the same as

∂cTab = Rkc (^
e i)a (^
e j)b Qijk (J.9.14a)

or, using (J.1.3) that (∇T)ijk ≡ ∂kTij,

(∇T)abc = Rkc (^
e i)a (^
e j)b Qijk . (J.9.15)

Our coefficients of interest are the (∇T)'abc which are related to (∇T)abc as in (E.8.20),

(∇T)'abc = MaAMbBMcC (∇T)ABC

= MaAMbBMcC { RkC (^
e i)A (^
e j)B Qijk } . (J.9.16)

Things will simplify. First, from (E.8.12) we know that (^

e i)A = NAi and (^
e j)B = NBj , so we continue
the above

(∇T)'abc = MaAMbBMcCRkCNAiNBj Qijk

= (MaANAi)(MbBNBj)McCRkC Qijk // reorder

= (δai)(δbj)McCRkC Qijk = McCRkC Qabk // M and N are inverses as in (E.8.9)

= (h'c RcC)RkC Qabk = h'c(RcCRkC)Qabk // (E.8.7) then reorder

= h'c(g'cdRdCRkC)Qabk = h'cg'cd(RdCRkC)Qabk // (7.5.9) with g = 1 then reorder

= h'c(g'cdδdk)Qabk = h'c(g'ck)Qabk // (7.6.4) orthog rule #4

= h'c(h'k-2 δc,k)Qabk = h'c(h'c-2)Qabc // (5.11.9) for orthog coords

= h'c-1Qabc

= h'c-1 [∂'cT'ab + T'sb^

Γ'acs + T'as^
Γ'bcs ] . // (J.9.10) (J.9.17)

The final result with the Lai affine connection is then,

383
 Appendix J: Expansion of (∇T)

(∇T)'abc = (∇T)'abc = h'c-1 [∂'cT'ab + T'sb^

Γ'acs + T'as^
Γ'bcs ] . (J.9.18)

Now do an index shuffle abc→ ijm to get,

(∇T)'ijm = h'm-1 [∂'mT'ij + T'sj^

Γ'ims + T'is^
Γ'jms ] . (J.9.19)

Expressing this result for Picture C1 of (F.1.1) removes the primes to give

(∇T)ijm = (∇T)ijm = hm-1 [∂mTij + T sj^

Γims + Tis^
Γjms ] . (J.9.20)

The corresponding equations (J.9.4) and (J.9.5) become in Picture C1 (see (F.1.1) and following text)

(∂j^ ^k ^ (d^ ^k (q
qn)i = Γ jn ^k) dx
i j
q n) = Γ jn qk ⇒ (J.9.4)
q = ^
d^ n Γk ^
jnq dxj
k
^k = ^ qi = ^
jn q • (∂j^ d^ Γkji ^
k
Γ qn) (J.9.5) qk dxj . (J.9.21)

where qn(x) are the tangent base vectors in the ξ-space of Picture C1 and ^
qn = qn(x)/hn.

We now translate these results to the Lai et al. notation as follows:

Us Lai
^
qn en // unit vectors in ξ-space
Tij Tij
^
Γi Γsmi // the Lai affine connection, note reverse index order
ms
ijm
(∇T) Mijm
∇T M . (J.9.22)

The translation of (J.9.20) and the right side of (J.9.21) is then

(∇T)ijm hm = [∂mTij + T sj^

Γims + Tis^
Γjms ] // us, now take s→q
Mijm hm = [∂mTij + TqjΓqmi + TiqΓqmj ] // Lai p 504 (8A.26) (J.9.23)

qi = ^
d^ Γkji dxj ^
qk // us
dei = Γijk dxj ek . // Lai p 502 (8A.12) (J.9.24)

Lai et al. use (J.9.23) to compute the (∇T)ijm for cylindrical and spherical coordinates on p 504 and 505.
Their results are in agreement with our calculations (J.8.1) and (J.7.9).

384
 Appendix K: Deformation Tensors

Appendix K: Deformation Tensors in Continuum Mechanics

Tensor-like objects appear everywhere in continuum mechanics. As was noted below (2.12.3),
continuum mechanics texts generally refer to all objects having indices as being "tensors", whether or not
these objects actually transform as tensors with respect to some underlying transformation. The most
commonly appearing tensors have two indices and are just 3x3 matrices associated with 3D space. In this
category, there are several kinds of stress tensors, and many kinds of strain and deformation tensors which
describe how a tiny volume of continuous matter (perhaps a tiny cube near some point x) changes shape
in response to some applied stress. For a fluid, a fixed stress pattern can cause a continuous ongoing
change of shape which is measured then by a "rate of deformation tensor" often called D.
An equation relating stress to strain/deformation is called a constitutive equation and describes the
"response" of some physical system to "stimulus". The constitutive equation for a spring is F = -kΔx, for
example, which is distinct from the equation of motion for a mass on a spring which is F = ma. For the
spring, the stimulus is the force F ("stress"), and the response is the spring stretch Δx ("strain").

In this Appendix we shall study the tensor aspects of several kinds of deformation tensors appearing in
continuum mechanics. In the book of Lai et al. this material is spread over several chapters, but here it
will all be put in one place with Lai references provided. Section K.3 below considers the form of a
candidate constitutive equation for a continuous solid whose form is determined by the requirement that
the equation be "covariant" as discussed in Section 7.15. Similarly, Sections K.4 and K.5 consider
covariant constitutive equations for fluids

K.1 A Preliminary Deformation Flow Picture

Our flow version of Picture A was shown as (5.16.1), where F ≡ F-1 :

(K.1.1)

The upper arrow represents an underlying generally non-linear transformation between x-space and x'-
space given by x' = F(x) while matrix S is the linearized-at-point-x version of the transformation which
defines the notion of a vector with dx = S dx' as in (2.1.6). Since F will have another meaning below, we
change the transformation name so that x = F(X) where F ≡ F-1 .

Warning: The x' in Fig (K.1.2) below is unrelated to the x' appearing in (K.1.1) which is really X. Also,
overloaded symbol S will be used both for reference frames and for linearized transformation matrices.

Consider now the picture below in which appear two sequential transformations Fto and Ft. There are
three spaces called X-space on the bottom, x-space in the middle, and x'-space on the top. The spaces are
associated with frames of reference S0, S and S'. Each frame has some set of basis vectors to be discussed

385
 Appendix K: Deformation Tensors

below. The linearized S matrices associated with the two transformations are shown to the right and are
given the names St0 = F on the bottom and St= Ft on the top. The transformation x' = Ft(x,τ) is a spatial
coordinate transformation only, the time coordinate τ is a parameter. In the picture below, time increases
in the upward direction, so τ > t > t0.

(K.1.2)

Consider for the moment just the bottom transformation. The transformation Fto ≡ F describes the
deformation of a particle of continuous matter which starts at position X and time t0 and ends up at
position x at time t. If we look at a large cube of continuous matter, we might find that it deforms in a
very complicated manner as determined by the non-linear transformation F applied to all the particles
within this large cube. The cube gets stirred up and is probably no longer recognizable. But if instead we
consider a differentially small starting cube at X and t0, we shall find that at time t that cube is at location
x but has been transformed into a tiny rotated parallelepiped whose axes are no longer orthogonal. It is
assumed that the flow is reasonable and smooth, we are not considering some kind of singular
"explosion" here. We use the words flow and fluid, but the deformation concept applies to elastic solids
as well as fluids since these deform in some way when they are stressed (think jello or even steel).

Rather than think of the flow in terms of the edges of this tiny cube, one can instead consider two
very closely spaced points in the fluid close to X which are separated by spacing dX at time t0, which we
think of as "a little dumbbell". At time t, if one carefully tracks the "pathlines" of the ends of the
dumbbell, one finds that the dumbbell tumbles and stretches and ends up as dx at time t and location x,

(K.1.3)

386
 Appendix K: Deformation Tensors

This differential dumbbell can be regarded as a mathematical "probe" embedded in the continuous
medium. The relationship between dx and dX is given by

dx = F(X,t) dX dxi = FijdXj // Lai p 105 (3.18.3) (K.1.4)

where the matrix Fij is called "the deformation gradient". Matrix F is also known as "the deformation
gradient tensor" even though it is not a "tensorial tensor" with respect to any identifiable transformation.
In Section 5.16 the above equation dx = FdX was identified with dx = Sdx' with S being F. Fig (K.1.3) is
just Fig (2.1.2) with x'-space identified with X-space. Thus, the deformation gradient F is the linearized
version (at point x) of some fancy non-linear (and unknown) "flow transformation" F. Since one can write

dxi = (∂xi/∂Xj)dXj , (K.1.5)

one finds from (K.1.4) that

Fij = (∂xi/∂Xj) ≡ ∂jXi or F = (∇x) // Lai p 105 (3.18.4) (K.1.6)

where the gradient ∇ is with respect to X, so it is really ∇ = ∇(X). Thus the name "deformation gradient"
for F. [ Notice that (∇x) is a matrix. In Appendix G the form of (∇v) for arbitrary vector v is found in
arbitrary curvilinear coordinates. The index order reversal Fij = ∂jXi is mentioned there as well. ]
The deformation gradient F(X,t) depends implicitly on the time t0. At t = t0+ε (with very small ε) no
flow has yet taken place, so dxi = dXi and then F(X,t0) = 1. Time t0 is called the reference time and one
could display it by writing F(X,t) = Ft0(X,t), but normally this t0 label is suppressed.

Now consider the upper flow in Fig (K.1.2) above. It is entirely analogous to the lower flow, but names
are changed. One gets from the lower flow to the upper flow by making these replacements :

(t, x, dx, S) → (τ, x', dx', S')

then
(t0, X, dX, S0) → (t, x, dx, S) . (K.1.7)

The equations corresponding to (K.1.4) are therefore (as shown in (K.1.2)),

dx' = Ft(x,τ) dx dx'i = (Ft)ijdxj . // Lai p 457 (8.7.2) (K.1.8)

Since one can write

dx'i = (∂x'i/∂xj)dxj ,

one finds that

(Ft)ij = (∂x'i/∂xj) or Ft = (∇x') // Lai p 457 (8.7.3) (K.1.9)

where the gradient ∇ is with respect to x, so it is really ∇ = ∇(x).

387
 Appendix K: Deformation Tensors

In the lower flow, t0 is the reference time, and t is the "current time". In the upper flow, the current time t
is the reference time, and τ is some time τ > t. The upper flow is relative to current time t as reference,
and for that reason the word relative is pre-pended to the names of all related tensors. Thus, Ft is called
the "relative deformation gradient ", whereas F is just the "deformation gradient ".

Why are relative tensors useful?

The main motivation for use of the relative tensors concerns differentiation with respect to time in the
vicinity of the current time t. One can write

dtFt(x,t) ≡ [∂τFt(x,τ)]τ=t // x fixed (for example, x = x1)

where
∂τFt(x,τ) ≈ [ Ft(x,τ+dτ) - Ft(x,τ) ] / dτ . (K.1.10)

Here dt = Dt = d/dt = D/Dt = the total time derivative, and ∂t = ∂/∂t = the partial time derivative. Since x
is fixed, dx = 0 so dt = ∂t. We want to know the rate of deformation at some fixed current time t, and it
is the τ argument of the function Ft(x,τ) that lets this derivative be computed.
One could in theory carry out this same differentiation using the "non-relative" tensors by doing d/dt0
with t0 near t:

dtFt(X,t) ≡ [∂t0Ft0(X,t)]t0=t // X fixed

where
∂t0Ft0(X,t) ≈ [ Ft0+dt0(X,t) - Ft0(X,t) ] / dt0 , (K.1.11)

but this goes against the grain of the idea that X is a material coordinate at constant, fixed initial time t0
which is earlier than t. And in the above, we end up with a statement about Lagrangian functions f(X,t)
rather than Eulerian functions f(x,t), though one might argue that as t0→ t, one has X → x. The relative
tensor approach just makes the differentiation process clearer, and will be used below for that purpose.
[ In the Lagrangian or material picture, one tracks a blob that started at position X -- one's
instrumentation so to speak flows with the blob. In the Eulerian picture the instrumentation is fixed in
space and observes the flow passing by. ]

The frames of reference and the cameraman

Each of the three frames of reference S0, S and S' in Fig (K.1.2) has its own set of orthonormal basis
vectors which we are completely free to set in any manner. As a construct it is helpful to imagine that, as
the flow proceeds, it is observed by a cameraman who flies around on a camera platform which translates
and rotates in some arbitrary manner. Since our main concern will be with the dumbbells like dX, dx and
dx', the translational part of the camera platform motion is irrelevant since dX is invariant under
translations. We allow the cameraman's arbitrary orientation at times t0, t and τ to determine the axes of
the three frames S0, S and S'. The cameraman is an "observer".
The values of the deformation gradient matrix elements Fij depend on the choice of basis vectors in
frames S0 and S, so they in fact are dependent on how the cameraman flies his platform. Consider,

388
 Appendix K: Deformation Tensors

dx = dx1^
e 1(S) + dx2 ^
e 2(S) + dx3 ^
e 3(S) (K.1.12)

dX = dX1^
e 1(S0) + dX2 ^
e 2(S0) + dX3 ^
e 3(S0) . (K.1.13)

Once the axes are chosen, the value of the Fij are determined, for example,

F12 ≈ (dx1)/(dX2) . (K.1.14)

If we were to rotate the basis vectors in frame S, for example, dx1 would change, dX2 would stay the
same, and F12 would change.

Tensor expansions of the deformation gradients

These two expansions won't be used below, they are just inserted here as an application of the work of
Appendix E on tensor expansions. These expansions use the cameraman basis vectors just defined above
which are ^
e n(S0) for frame S0 and ^
e n(S) for frame S.

• Consider this candidate expansion for the deformation gradient F,

F = Σij Fij ^
e i(S) ⊗ ^
e j(S0) = Σij Fij [^
e i(S)] [^
e j(S0)]T

where Fij = [F(S,S0)]ij = [^

e i(S)]T F [^
e j(S0)] = (∂xi/∂Xj) , (K.1.15)

where we write the expansion in both direct product and matrix forms of Appendix E. This is a "mixed
basis expansion" as discussed in Section E.10. Consider the application of this expansion to dX :

{ Σij Fij [^
e i(S)] [^
e j(S0)]T } dX // ≡ {expansion} dX

= Σij Fij [^
e i(S)] [^
e j(S0)]T { ΣkdXk^
e k(S0)}

= Σij Fij ΣkdXk [^

e i(S)] [^
e j(S0)]T [ ^
e k(S0)]

= Σij Fij ΣkdXk [^

e i(S)] δj,k

= [ Σij Fij dXj] ^

e i(S)

= [ dxi ] ^
e i(S) // using the fact that F dX = dx

= dx . (K.1.16)

Since {expansion}dX = dx and since (∇(X)x) dX = dx by the chain rule, it seems reasonable to conclude
that {expansion} = (∇(X)x) = F.

389
 Appendix K: Deformation Tensors

• If we agree to use the ^

e i(S) for both dx and dx', so that dx = dxi^ e i(S) and dx' = dx'i^
e i(S), then the
following is a viable expansion for the relative deformation gradient Ft:

Ft = Σij (Ft)ij ^
e i(S) ⊗ ^
e j(S) = Σij (Ft)ij [^
e i(S)] [^
e j(S)]T

where (Ft)ij = [Ft(S,S)]ij = [^

e i(S)]T Ft [^
e j(S)] = (∂x'i/∂xj) (K.1.17)

The verification is similar to the above,

{ Σij (Ft)ij [^
e i(S)] [^
e j(S)]T } dx // ≡ {expansion} dX

= Σij (Ft)ij [^
e i(S)] [^
e j(S)]T { Σkdxk^
e k(S)}

= Σij (Ft)ij Σkdxk [^

e i(S)] [^
e j(S)]T [ ^
e k(S)]

= Σij (Ft)ij Σkdxk [^

e i(S)] δj,k

= [ Σij (Ft)ij dxj] ^

e i(S)

= [ dx'i ] ^
e i(S) // using the fact that Ft dx = dx'

= dx' . (K.1.18)

Since {expansion}dx = dx' and since (∇(x)x') dx = dx' by the chain rule, it seems reasonable to conclude
that {expansion} = (∇(x)x') = Ft.

390
 Appendix K: Deformation Tensors

K.2 A More Complicated Deformation Flow Picture

Consider now this flow picture,

(K.2.1)

There is much to be said about this drawing.

The left side is the same as in Fig (K.1.2) shown above.
The picture is simplified in that it shows only the linearized S-type transformations like F and not the
full transformations like F, and these S-type matrices are now labeled right on the transformation arrows.
We only care about these F matrices because we only care about the dumbbells like dx. For example, on
the lower left we have dx = F dX .
The two sides of the picture represent observations of the same flow by two independent flying
cameraman observers, call them C and C*. On each side the basis vectors of the various frames are set by
the motions of these cameramen. The two cameramen have agreed to start off at time t0 with their camera
platforms in exact alignment, so there is no need for a frame S0*.
The two frames of reference S and S* are related by some Galilean transformation (rotation plus
translation) which brings the two independent camera platforms into alignment at time t :

x* = Q(t) (x-x0) + c(t) ⇒ dx* = Q(t) dx . (K.2.2)

Here x0 is a randomly selected origin for the rotation Q(t), and c(t) the corresponding translation. As
above, we only care about the Q(t) part of this transformation, so in terms of dx objects, the frames S and
S* are in effect related by the rotation Q(t).
The arrows in Fig (K.2.1) correctly describe the transformations of the dx type objects in moving
between frames. In the lower part, for example, we have

dx* = Q(t) dx dx = F dX dx* = F* dX . (K.2.3)

391
 Appendix K: Deformation Tensors

Comparing the left and right equations one has Q(t) dx = F* dX and then the center equation can be used
on the left side to get Q(t) F dX = F*dX . Since this has to be true for any dX, we have Q(t) F = F* as
shown in the drawing. This result is trivially obtained just by looking at the alternate arrow paths from
frame S0 to frame S*. So:

F* = Q(t) F or F*(x*,t*) = Q(t) F(x,t) . t* = t (K.2.4)

At time τ we have a similar situation, but there are four arrows instead of three. Comparing the arrow
paths from frame S to frame S'* one finds

Ft*Q(t) = Q(τ)Ft ⇒

Ft* = Q(τ)FtQ(t)T or Ft*(x*,τ*) = Q(τ) Ft(x,τ)Q(t)T t* = t (K.2.5)

The two Q's are rotations (reflections included) and are therefore orthogonal so Q-1 = QT.

Does F transform as a tensor with respect to rotation Q(t) ?

We can think of F(x,t) as a property of the continuous material at location x and current time t. F
describes the "state of deformation". If F transformed as a tensor with respect to Q(t), one would need this
to be true,

F* = Q(t) F Q(t)T , // not true! (K.2.6)

which is the matrix form for the transformation of a rank-2 tensor as shown in (5.7.3). But we have just
seen that F* = Q(t) F so the required Q(t)T on the right is missing. We conclude therefore that in fact F,
although it is called a tensor, does not transform as a tensor under Q(t). One then says that F is a non-
objective tensor with respect to Q(t). Equation F* = Q(t) F in fact says that the columns of matrix F
transform as vectors under Q(t), which is very different from saying F transforms as a rank-2 tensor under
Q(t).
If one is trying to construct a phenomenological equation modeling a continuous material at point x
and time t, one must make sure that equation is "covariant" (frame-indifferent) with respect to rotation
Q(t). The observers (cameramen) in frame S and frame S* must see equations which have exactly the
same form, which means the elements in the equations must be objective with respect to Q(t). See Section
7.15 for a general discussion of "covariance". Since F is non-objective, it is not directly useful in the
construction of covariant model equations.

Do any of the usual "derived tensors" transform as tensors with respect to Q(t) ?

By "the usual derived tensors" we mean B, C, U, V and associated R all defined as follows:

B = FFT = the left Cauchy-Green deformation tensor = the Piola deformation tensor (K.2.7)

C = FTF = the right Cauchy-Green deformation tensor = the Finger deformation tensor (K.2.8)

F = RU = VR R = rotation U,V = symmetric positive definite (K.2.9)

392
 Appendix K: Deformation Tensors

Tensors B and C are defined as shown, and both are therefore symmetric tensors. The last line is a
statement of the polar decomposition theorem which says that any (real) non-singular matrix (det ≠ 0)
can be uniquely written in these two ways (we apply this theorem to the deformation tensor F)

F = RU = VR ⇒ U = RTVR and V = RURT // Lai p 110 (3.21.1,2,4) (K.2.10)

where R is a rotation matrix and V and U are symmetric positive definite matrices (meaning the
eigenvalues are all positive) known as the left and right stretch tensors. Note that R is the same matrix in
both the RU and VR forms. The idea is that the R matrix takes into account the rotational part of the
deformation F, while U or V take into account the stretch component of the deformation. If the
deformation is a pure rotation, U = V = 1, whereas if the deformation is a pure stretch then R = 1. A
general deformation is a rotation/stretch/shear affair and one will find that none of R, U, V are unity.
One can combine the three equations above to find that

B = FFT = (VR)(VR)T = VRRTVT = VVT = VV = V2 // Lai p 121 (3.25.1) (K.2.11)

C = FTF = (RU)T(RU) = UTRTRU = UTU = U2 . // Lai p 115 (3.23.1,2) (K.2.12)

So our task is to discover whether any of these derived tensors transform as a tensor relative to Q(t). If
they do transform as tensors (if they are objective), then they are candidates for use in constructing model
equations for the continuous material.
We start with B and C:

B* = F*F*T = (QF)(QF)T = QF FTQT = QBQT ⇒ B* = Q(t)BQ(t)T (K.2.13)

C* = F*TF* = (QF)T(QF) = FTQTQF = FTF = C ⇒ C* = C . (K.2.14)

Thus, the left Cauchy-Green deformation tensor B actually does transform as a rank-2 tensor with respect
to Q(t), so it is a tensorial tensor, it is "objective". In contrast, since C* = C, the right Cauchy-Green
deformation tensor does not transform as a rank-2 tensor. In fact each element of matrix C transforms as a
tensorial scalar with respect to Q(t).
What about V and U as defined above, the left and right stretch tensors?

F = RU = VR F* = R*U* = V*R* . (K.2.15)

Consider,

F* = QF = Q(RU) = (QR) (U) = R*U* . (K.2.16)

Since U is positive definite symmetric, and since QR is a rotation, and since the polar decomposition is
unique, it must be that

R* = QR and U* = U . (K.2.17)

Next write

393
 Appendix K: Deformation Tensors

F* = QF = Q(VR) = (QVQT)(QR) = V* R* . (K.2.18)

Since the eigenvalues of symmetric V are determined by det(V-λI) = 0, and since this is the same as the
equation det(QVQT-λI) = 0, QVQT has the same eigenvalues as V and so (QVQT) is symmetric and
positive definite. Due to this fact and the fact that QR is a rotation, and the fact that the polar
decomposition is unique, it must be that

V* = QVQT and QR = R* . (K.2.19)

and thus V transforms as a true tensor. So here is a summary for our tensors of interest. Only two of the
five deformation tensors actually transform as tensors. The references are to Lai page 336-337 :

F* = Q(t)F // Lai (5.56.21)

B* = Q(t)BQ(t)T // rank-2 tensor with respect to Q(t) so objective // Lai (5.56.31)
C* = C // Lai (5.56.28)
U* = U
V* = Q(t)VQ(t)T // rank-2 tensor with respect to Q(t) so objective
R* = Q(t)R (K.2.20)

Comment: Recall that F = F(x,t) has a hidden parameter t0 so in fact F = Ft0(x,t). Similarly, all derived
tensors have this same hidden parameter. Thus, for example, one could write the transformation of B as

Bt0*(x*,t*) = Q(t) Bt0(x,t)Q(t)T t* = t x* = Q(t) (x-x0) + c(t) dx* = Q(t) dx . (K.2.21)

The parameter t0 is treated as a fixed constant here and plays no role in the question of whether or not B
transforms as a rank-2 tensor. The important time argument of B is the current time t, and the main idea is
that B*(t) = Q(t) B(t)Q(t)T so that B(t) is objective with respect to the rotation Q(t). The transformation is
valid for any value of t0. In the limit that t0 → t, the equation says 1 = Q(t) 1 Q(t)T which of course is true
since rotation Q(t) is orthogonal.

Do any of the usual relative derived tensors transform as tensors with respect to Q(t) ?

Again we think of a relative tensor Wt as being a property of the continuous material at current time t, a
measure of the state of deformation. Such a tensor is objective only if Wt* = Q(t)WtQ(t)T. With regard to
the above Comment, in this new situation it is the t of Wt which is the time variable of interest (the
current time), and time τ is regarded as a fixed parameter, as was t0 in the Comment. It just happens that
the notational positions of the current time t and the parameter time τ are swapped in this case relative to
the last, so now we have

Wt*(x*,τ*) = Q(t)Wt(x,τ) Q(t)T τ* = τ x* = Q(t) (x-x0) + c(t) dx* = Q(t) dx . (K.2.22)

394
 Appendix K: Deformation Tensors

A tensor Wt which transforms as a rank-2 tensor (is objective) with respect to rotation Q(t) must satisfy
the relation above, where the arguments of both Q rotations are t. As in the Comment above, this
transformation is valid for any value of parameter τ, and as τ→t, the equation says 1 = Q(t) 1 Q(t)T.

Our study of the transformation properties of the relative tensors proceeds in a manner similar to that used
for the regular tensors above. We start with Bt ≡ FtFtT :

Bt* = Ft*Ft*T = [Q(τ) Ft QT(t)] [Q(τ) Ft QT(t)]T = Q(τ) Ft QT(t) Q(t) FtT Q(τ)T

= Q(τ) Ft FtT Q(τ)T = Q(τ) Bt Q(τ)T // not a rank-2 tensor since t ≠ τ (K.2.23)

Next comes Ct ≡ FtTFt :

Ct* = Ft*TFt* = [Q(τ) Ft QT(t)]T [Q(τ) Ft QT(t)] = Q(t) FtT Q(τ)T Q(τ) Ft QT(t)

= Q(t) FtT Ft QT(t) = Q(t) CtQT(t) // yes a rank-2 tensor with respect to Q(t) (K.2.24)

What about the left and right relative stretch tensors Vt and Ut?

Ft= RtUt = VtRt Ft* = Rt*Ut* = Vt*Rt* (K.2.25)

Consider,

Ft* = Q(τ) Ft QT(t) = Q(τ) RtUtQT(t) = [Q(τ) RtQT(t)] [Q(t)UtQT(t)] = Rt*Ut* . (K.2.26)

Since [Q(τ) RtQT(t)] is a rotation and since [Q(t)UtQT(t)] is a symmetric positive definite matrix by the
argument given in the previous section, and since the polar decomposition is unique, it must be that

Rt* = Q(τ) RtQT(t) and Ut* = Q(t)UtQT(t) // Ut is a rank-2 tensor (K.2.27)

Finally, write

Ft* = Q(τ) FtQT(t) = Q(τ)VtRtQT(t) = [Q(τ)VtQT(τ)] [Q(τ) RtQT(t)] = Vt* Rt* (K.2.28)

By the same argument used several times above, we conclude that

Rt* = Q(τ)RtQT(t) and Vt* = Q(τ)VtQT(τ) // Vt is not a rank-2 tensor, τ ≠ t (K.2.29)

The rule for transforming Rt is the same as found a few lines above.

395
 Appendix K: Deformation Tensors

Here then are the conclusions, with references to Lai page 472:

Ft* = Q(τ)FtQT(t) // Lai (8.13.6)

Bt* = Q(τ)BtQ(τ)T // Lai (8.13.12)
Ct* = Q(t)CtQT(t) // rank-2 tensor with respect to Q(t) so objective // Lai (8.13.10)
Ut* = Q(t)UtQT(t) // rank-2 tensor with respect to Q(t) so objective // Lai (8.13.9)
Vt* = Q(τ)VtQT(τ) // Lai (8.13.12)
Rt* = Q(τ)RtQT(t) // Lai (8.13.8) (K.2.30)

Notice that among the "normal" tensors, B and V are objective, whereas among the "relative tensors" it is
Ct and Ut that are objective. All the other tensors are "non-objective".

K.3 Covariant form of a solid constitutive equation involving the deformation tensor

For a solid continuous material in frame S one can consider a stress/deformation relationship of the form
T = f(B), where T is the Cauchy stress tensor, B is the left Cauchy-Green deformation tensor mentioned
in (K.2.7) above, and f is "some function".
In frame S*, there will be some covariant version of the equation T* = f*(B*). If the medium is
isotropic (rotationally invariant in its properties), then f* = f and one will have T* = f(B*) in Frame S*.
Two observers of the same system in frames related by a rotation cannot observe different functions f ≠ f*
if the material is isotropic. Notice that there are two separate issues here: (1) equation must be covariant
under rotations to be viable; (2) isotropic implies f = f*.
If f is a polynomial, or a function which can be approximated by one (f is smooth), then T = f(B) with
polynomial coefficients which are rotational scalars (with respect to Q) is a viable equation form for the
following reason: since B is a rank-2 tensor by (K.2.20), so is any power of B,

B*2 = [QBQT][QBQT] = Q B2QT etc. (K.3.1)

and if the polynomial coefficients are scalars, then f(B) is a rank-2 tensor.
Just as a particle force F transforms as a rank-1 tensor under rotations, the Cauchy stress tensor T
transforms as a rank-2 tensor under rotations, and then both sides of T = f(B) transform in the same way --
as rank-2 tensors. Any candidate equation between T and a deformation tensor which did not have both
sides transforming the same way would be invalid from the get-go (except perhaps as an approximation).
The scalar coefficients must be functions of the Bij and there are three such scalars known as the
principal scalar invariants of B (Lai p 40), one of which is det(B), so the scalar coefficients can be any
functions of these three scalar invariants. Furthermore, one can use the fact that B = FFT is symmetric
along with the Cayley-Hamilton theorem (symmetric matrix B satisfies its own secular equation, whose
coefficients by the way are those scalar invariants) to show that any powers of B in polynomial f(B) larger
than degree 2 can be expressed as a linear combination of I, B and B2. One ends up then with T = aI + bB
+ cB2 where a,b,c are functions of the three scalar invariants of tensor B.
Since both sides of T = f(B) transform in the same way (rank-2 tensors), the equation T = f(B) is
"covariant" as discussed in Section 7.15, meaning it has the same form in frame S* as it has in S.
The equation T = f(B) is a relation between stress and strain in the form of deformation, and as such
is called a constitutive equation for the continuous material. One wants such equations to be covariant
between frames of reference related by any Galilean transformation (rotation + translation), even if one or

396
 Appendix K: Deformation Tensors

both of these frames are non-inertial. This is an extension of Hooke's Law for a spring, F = -k Δx , which
is covariant under rotations and translations.
In contrast, equations of motion are covariant only if both frame S and S* are inertial frames.
Notice that this entire discussion falls apart completely if one tries T = f(F) or T = f(C) as a candidate
constitutive relation, since then the two sides of the equation don't transform the same way.
This subject is discussed in Lai pp 334-342 and p 40 for the scalar invariants. The requirement of
covariance for an isotropic material and the fact that B is symmetric and transforms as a tensor puts a
severe restriction on the form of the constitutive equation and we end up with T = aI + bB + cB2. Since
one can replace B3 = αB2 + βB + γI, if B is invertible (detB ≠ 0) one has B2 = αB + βI + γB-1 and this
allows the alternate form T = a'I + b'B + c'B-1 . This last equation is used to model large deformations of
an isotropic elastic material. An example is the Mooney-Rivlin theory for rubber.

K.4 Some fluid constitutive equations

It was noted just above (K.1.10) that the relative deformation tensors are appropriate when one is
interested in time derivatives of the tensors. It was also noted in (K.2.30) that the relative deformation
tensor Ct is objective. One can expand Ct(x,τ) in a Taylor series about current time t in this manner (∂τ
≡ ∂/∂τ) ,

∞ ∞
Ct(x,τ) = Σn=0 [ ∂τnCt(x,τ)]τ=t (τ-t)n/n! = Σn=0 An(x,t) (τ-t)n/n! // Lai p 463 (8.10.1)

An(x,t) ≡ [ ∂τnCt(x,τ)]τ=t , (K.4.1)

where the coefficient derivatives are given the names An(x,t) called Rivlin-Ericksen tensors. Each of
these coefficient tensors is in fact objective, just as is Ct, since (as usual, t = t*, τ = τ* )

Q(t) [ ∂τnCt(x,τ)]τ=t QT(t) = { ∂τn [Q(t) Ct(x,τ) QT(t)]}τ=t = { ∂τ*n Ct*(x*,τ*)}τ*=t

⇒ Q(t) An(x,t) QT(t) = A*n(x*,t) (K.4.2)

These An(x,t) tensors appear in various covariant models of "non-Newtonian" fluid behavior, the general
study of which is called rheology, based on the Greek word for a current flow (a rheostat controls electric
current),

// OED2

Here are a few covariant constitutive equations and the names assigned to them (Lai p 481). Note that for
any normal fluid, there is always a -pI tensor term in the expression for stress T, where p is the fluid
pressure and I is the identity matrix. The diagonal elements of matrix -pI are the equal normal stresses of
the surroundings of a tiny cube of fluid pulling out on the cube faces, hence the -p (p > 0) since we know
the fluid actually pushes in on the cube.

397
 Appendix K: Deformation Tensors

T = -pI + functional of Ct(τ), τ ≤ t // "simple" fluid, since ∇nFt not involved (Ct=FtT Ft)

T = -pI + ∫
t
dτ f1(τ) Ct(τ) // single-integral simple fluid. f1(τ) = a memory weight function
-∞

T = -pI + f(A1, A2....AN) // Rivlin-Ericksen incompressible fluid of complexity N

T = -pI + f(A1,A2) // viscometric flow fluid (there are conditions on A1 and A2)
T = -pI + μ1A1 + μ2A12 + μ3A2 // second order fluid (paint, blood, polymers)

T = -pI + μA1 // incompressible Newtonian fluid (fluids like water) (K.4.3)

It turns out that A1 = 2D where D = [(∇v) + (∇v)T] /2 ≡ (∇v)sym , so A1 is twice the rate of deformation
tensor D. The other An can then be found from this recursion relation,

An+1 = dtAn + An(∇v) + (∇v)TAn // Lai p 468 (8.11.2) (K.4.4)

Here v is the fluid velocity vector and dt = d/dt = D/Dt. Again, (∇v) is the subject of Appendix G.

K.5 Corotational and other objective time derivatives of the Cauchy stress tensor

The Cauchy stress tensor T transforms as a tensor under Q(t); it is objective. One can write therefore,

T*(x*,t*) = Q(t) T(x,t) Q(t)T t* = t x* = Q(t) (x-x0) + c(t) dx* = Q(t) dx . (K.5.1)

Clarification of the above equation

One can think of the above equation T* = QTQT as involving operators in Hilbert Space, as outlined in
Section E.7. In the upper part of Fig (K.2.1) above we show four different frames of reference called S,
S*, S' and S'* each of which has its own set of basis vectors we might call un, u*n, u'n and u*'n. It
happens that the picture refers to S and S* at time t, and S' and S'* at time τ, but any basis vectors can be
"used" at any time one wants. For example, here are four expansions of the operator T(x,t)

T(x,t) = Σab Tab(x,t) ua ⊗ ub = Σab T*ab(x*,t) u*a ⊗ u*b

= Σab T'ab(x',t) u'a ⊗ u'b = Σab T'*ab(x'*,t) u'*a ⊗ u'*b (K.5.2)

in which we see four different kinds of components Tab, T*ab, T'ab, T'*ab . The spatial arguments of
each component are written as appropriate for that frame of reference and of course all "correspond" to
each other (for example, x' = Ft(x,τ)). Recall from (2.5.1) the notion of the transformation of a
contravariant vector field in developmental notation

V'(x') = R V(x) contravariant Rik(x) ≡ (∂x'i/∂xk) R = S-1 (2.5.1)

where the argument is appropriate to the space of interest.

398
 Appendix K: Deformation Tensors

The time argument t in the above four expansions of T can be set to any arbitrary value. The stress
tensor at a point x is in general a function of time t. One could for example set t = τ in all the expansions.
Having said this, we now decide that only the frame S basis vectors un shall be used in our
expansions and components. Then for example ( these un were called ^ e n(S) earlier)

T(x,t) = Σab Tab(x,t) ua ⊗ ub

T*(x*,t) = Σab T*ab(x*,t) ua ⊗ ub

Q(t) = Σab Qab(t) ua ⊗ ub . (K.5.3)

Our operator statement of objectivity then becomes the following when expressed in components,

T*(x*,t*)ij = Q(t)ia T(x,t)ab Q(t)Tbj t* = t . (K.5.4)

Thus, there should be no confusion about the following two equations which we express back in operator
notation with the position arguments suppressed (but shown on the right)

T*(t) = Q(t) T(t) Q(t)T // T*(x*, t) = Q(t) T(x,t) Q(t)T

T*(τ) = Q(τ) T(τ) Q(τ)T // T*(x*, τ) = Q(τ) T(x,τ) Q(τ)T . (K.5.5)

Problem: The tensor dT/dt fails to transform as a rank-2 tensor, even though T does so transform.

If one tries to construct covariant constitutive equations involving dT/dt, a problem arises because dT/dt is
non-objective,

T*(t) = Q(t) T(t) Q(t)T

(dT*/dt) = Q (dT/dt) QT + [ (dQ/dt) T QT + Q T (dQ/dt)T ] , (K.5.6)

so there are two extra unwanted terms. Just as B = FFT is constructed to provide an objective derived
tensor from non-objective F, one can construct a derived version of (dT/dt) which is objective. In the next
three sections, three different derived versions are described.

The corotational/Jaumann derivatives

The first step is to define an adjusted stress tensor Jt(τ) at time τ according to (see Lai p 483 (8.19.3). Lai
does not have a t subscript on J).

Jt(τ) ≡ RtT(τ) T(τ) Rt(τ) // Jt(x,τ) ≡ RtT(x,τ) T(x,τ) Rt(x,τ) (K.5.7)

where Rt(τ) is the rotation which appears above in (K.2.25), where we had (showing τ arguments),

399
 Appendix K: Deformation Tensors

Rt*(τ) = Q(τ) Rt(τ)QT(t) . (K.2.27) (K.5.8)

The tensor Rt(τ) is non-objective due to appearance of Q(τ) instead of Q(t) on the left (see comments near
(K.2.22) on Wt ). Recall that this rotation Rt(τ) is unique and is determined from the deformation tensor
by the polar decomposition Ft(τ) = Rt(τ)Ut(τ) = Vt(τ) Rt(τ). Thus, in some sense Jt(τ) knows about the
stress tensor T(τ), and it knows something about the deformation tensor through Rt(τ). [ The meaning of
the term "corotational" is explained far below. ]

The claim now is that the time derivative of this corotating stress tensor Jt is objective, meaning that
tensor dtJt transforms as a rank-2 tensor under the rotation Q(t). Here is a proof :

We first assemble the following facts,

T*(τ) = Q(τ) T(τ) Q(τ)T // transformation of stress tensor T at time τ, (K.5.5)

Rt*(τ) = Q(τ) Rt(τ)QT(t) // how Rt(τ) transforms, where Ft(τ) = Rt(τ)Ut(τ), (K.5.8)
Jt(τ) ≡ RtT(τ) T(τ) Rt(τ) // definition of Jt(τ) in frame S, (K.5.7)
Jt*(τ) ≡ Rt*T(τ) T*(τ) R*t(τ) // corresponding Jt* in frame S* (K.5.9)

and then we combine these ingredients to obtain a transformation rule for Jt :

Jt*(τ) ≡ Rt*T(τ) T*(τ) R*t(τ) = [Q(τ) Rt(τ)QT(t)]T [Q(τ) T(τ) Q(τ)T] [Q(τ) Rt(τ)QT(t)]

= [Q(t) RtT(τ)QT(τ)] [Q(τ) T(τ) Q(τ)T] [Q(τ) Rt(τ)QT(t)]

= Q(t) RtT(τ) [QT(τ)Q(τ)] T(τ) [Q(τ)TQ(τ)] Rt(τ)QT(t)

= Q(t) [ RtT(τ)T(τ) Rt(τ)] QT(t)

= Q(t) Jt(τ) QT(t) . (K.5.10)

Since this equation Jt*(τ) = Q(t) Jt(τ) QT(t) fulfills the condition described earlier for Wt to be objective,
we conclude that the corotating stress transforms as a rank-2 tensor, where τ is treated as a parameter.

Consider now the limit of (K.5.10) as τ → t. One finds,

Jt(τ) ≡ RtT(τ) T(τ) Rt(τ) // (K.5.7)

Jt(t) ≡ RtT(t) T(t) Rt(t) = 1 T(t) 1 = T(t) (K.5.11)
and
J*t(τ) ≡ R*tT(τ) T*(τ) R*t(τ) // (K.5.9)
T
J*t(t) ≡ R*t (t) T*(t) R*t(t) = 1 T*(t) 1 = T*(t) . (K.5.12)

In this limit, the corotation Rt-1(τ) has come to a halt, and (K.5.10) becomes a statement that T is
objective.

400
 Appendix K: Deformation Tensors

More interestingly, we can apply ∂τn = ∂n/∂τn to both sides of(K.5.10) to get

dτn Jt*(τ) = Q(t)[ dτn Jt(τ) ] QT(t) . (K.5.13)

Taking the limit τ→t then gives

[dtn Jt*](t) = Q(t) [dtn Jt](t) QT(t) (K.5.14)

which says that dtnJt are all objective tensors. And in particular, for n = 1,

(dtJt)* = Q(t) (dtJt) QT(t) , (K.5.15)

and this concludes our proof that dJt/dt is objective, whereas dT/dt is not objective.

The above objective tensor time derivatives are sometimes written using the following strange notation

o o o
Tn ≡ [dnJt(t)/dtn], n = 1,2,3... T ≡ T1 Jt(τ) ≡ RtT(τ) T(τ) Rt(τ) (K.5.16)

and these are called corotational or Jaumann derivatives (Lai p 484) [Jaumann-Zaremba]. It can be shown
that
o
T = dtT + TW-WT where W = [(∇v) – (∇v)T]/2 = "the spin tensor" // Lai p 484 (8.19.10)
(K.5.17)
The Oldroyd Lower convected derivatives

An alternative solution to the same problem uses a different adjusted stress tensor,

JL(τ) ≡ FtT(τ) T(τ) Ft(τ) // Lai p 484 (8.19.12) (K.5.18)

We suppress the t subscript on JL just to avoid having to write (JL)t(τ). In the table (K.2.30) one sees that
Ft transforms the same way Rt does, so one can repeat the above analysis to conclude that the derivatives
[dnJL(t)/dtn] are all objective tensors (just replace Rt→ Ft everywhere), so

Δ Δ Δ ∪
Tn ≡ [dnJL(t)/dtn] , n = 1,2,3... T ≡ T1, JL(τ) ≡ FtT(τ) T(τ) Ft(τ) // = Tn (K.5.19)

∪ Δ
and these are the "Oldroyd lower convected derivatives" (Lai p 485 uses Tn) . T is sometimes called the
Cotter-Rivlin stress rate. It can be shown that

Δ ∪
T = T = dtT + T(∇v) + (∇v)TT . // Lai p 485 (8.19.21) (K.5.20)

The Oldroyd Upper convected derivatives

Finally, consider again from (K.2.30) the non-objective way that Ft transforms,

401
 Appendix K: Deformation Tensors

Ft*(τ) = Q(τ)Ft(τ)QT(t)

⇒ (Ft-1)*(τ) = Q(t) (Ft-1(τ)) QT(τ) // inverted

⇒ (Ft-1)T*(τ) = Q(τ) (Ft-1)T(τ) QT(t) . // then transposed (K.5.21)

This object Ft-1,T therefore transforms the same way Rt and Ft transform, so we obtain a third set of
^ )
objective time derivatives called the Oldroyd upper convected derivatives (Lai p 486 uses Tn

∇ ∇ ∇
Tn ≡ [dnJU(t)/dtn] , n = 1,2,3... T ≡ T1, JU(τ) ≡ Ft-1(τ) T(τ) Ft-1,T(τ) ^
// = T (K.5.22)
n

The meaning of the term "convected" is explained below. It can be shown that

∇
^ = d T – (∇v)T – T(∇v)T
T=T // Lai p 486 (8.19.26) (K.5.23)
t

Covariant constitutive equations

Constitutive equations involving an objective time derivative of the stress tensor are called "rate type
constitutive equations". Here are some models for incompressible fluids :

o
T = -pI + S where S + λS = 2μD // a convected Maxwell fluid

T = -pI + S where S + λ(∂S/∂t) = 2μD // linear Maxwell fluid, see below (non-covariant)

T = -pI + S where S = 2μD // Newtonian fluid

o o
T = -pI + S where S + λ1S = 2μ(D + λ2D) // a corotational Jeffrey fluid

∇ ∇
T = -pI + S where S + λ1T = 2μ(D + λ2T) // Oldroyd fluid A (K.5.24)

Fluids with stress time derivatives in their constitutive equations exhibit both elastic and viscous behavior
at the same time. Pull on a chunk of such a fluid and the pull is initially resisted by an elastic force, but
after a while the internal stress field damps out (molasses, honey) and that elastic force goes away, as if
the fluid were microscopically constructed of little springs and dragging dashpots. When a constitutive
equation includes a time derivative of stress, the "response" (in this case D = [(∇v) + (∇v)T]/2 ) to the
"stimulus" (T or S) includes factors of the form e-t/c where the c are decay time constants which are
functions of the fluid parameters λi. In this case, the fluid has memory of its past over a time period less
than these time constants, as with the honey example. For flow that is very slow relative to these time
constants, the time derivative term may be neglected. In the moderately slow flow case, it can be shown
o
that the distinction between the corotational time derivative S and (dS/dt) can be neglected and then the

402
 Appendix K: Deformation Tensors

convected Maxwell fluid shown above becomes the traditional linear Maxwell fluid which is modeled on
those springs and dashpots with S + λ (∂S/∂t) = 2μD. (The time derivatives here are meant to act only on
the second argument of S(x,τ) so may be regarded as partial derivatives. ) If λ = 0, the linear Maxwell
fluid becomes an (incompressible) Newtonian fluid like water which has no memory.

Comment: The linear Maxwell fluid equation S + λ (∂S/∂t) = 2μD can be solved for S using the standard
∫-∞
t
Green's Function method and the solution is S(t) = 2 dt' [ (μ/λ)e-(t-t')/λ] D(t') where the bracketed
quantity (the Green's Function or kernel) is called the stress relaxation function φ(t-t'). One can see in this
solution the notion of memory (history) with time constant λ: the stress of the present is a function of the
rate of deformation D going on in the entire past history. This solution fits into the "simple fluid" form
shown earlier, where recall that D = (1/2)A1 and A1 = [ ∂τCt(x,τ)]τ=t.

Our main point is to demonstrate the construction of constitutive equations which are covariant with
respect to rotations, and which therefore can contain only tensors which in fact transform as tensors under
rotations. In continuum mechanics, such tensors are said to be objective tensors.

Interpretation of the adjusted stress tensors discussed above.

In Chapter 2 we discuss the notion of the transformation of a contravariant vector V' = RV in

developmental notation. In x'-space, the vector components are V'i = RijVi where Vi are the
components in x-space. If R is a rotation matrix, then the unit basis vectors in the two spaces can be taken
as Cartesian, call them u'n in x'-space and un in x-space. We have these two expansions of V:

V = Σn Vn un = Σn V'n u'n where Vn = V • un V'n = V • u'n . (K.5.25)

In the "active view" of things, we can think of V' = RV as creating a new vector V' in x-space from the
old vector V by rotating the vector V by R. In the "passive view", we think of the V'i as the components
of the original vector V projected onto the backwards-rotated basis vectors u'n = R-1 un. To verify this
relation between the basis vectors, we can write

V'n = V • u'n = V • R-1un = RV • RR-1un = RV • un = V'• un = V'n . (K.5.26)

So one can think either of V being rotated forward in x-space into V' where V' has x-space components
V'n , or one can think of the V'n as the components of V one measures in frame that is backwards rotated
by R-1 , that is, u'n = R-1 un.
Consider then a rank-2 tensor M that transforms as shown in (5.7.1) according to M' = R M RT. The
passive interpretation is that the components M'ij are those one observes in a frame of reference whose
basis vectors are rotated by R-1 relative to the basis vectors of the unprimed frame, just as in the vector
case of the last paragraph. If we now set R = R-1, then M' = R-1 M (R-1)T tells us that the components
M'ij of tensor M are those measured in a frame whose basis vectors are rotated forward by R relative to
the unprimed frame. If it happens that R = R(t), we would say that M'ij are the components of M which
are observed in a frame of reference which is rotating by R(t) relative to the frame of the unprimed
components Mij. The basis vectors of the primed frame are then u'n = R un .

403
 Appendix K: Deformation Tensors

With this long-winded introduction, we now consider the corotational stress tenser Jt(τ) shown in (K.5.7),

Jt(τ) ≡ RtT(τ) T(τ) Rt(τ) . (K.5.7) (K.5.27)

Since Rt(τ) is a rotation, RtT(τ) = Rt-1(τ), so we have, suppressing τ,

Jt = Rt-1T Rt . (K.5.28)

Therefore, we can regard (Jt)ij as T'ij, the components of stress T measured in a frame which is rotating
by Rt relative to the frame in which the Tij are measured. Since this primed frame rotates by Rt relative
to the unprimed frame, it is called a corotating frame, and Jt is then called the corotational stress, and its
time derivative is called the corotational stress rate.

We next consider the upper Oldroyd stress defined above in (K.5.22),

JU ≡ Ft-1 T (Ft-1)T . (K.5.22) (K.5.29)

In analogy with the above discussion, we can regard (JU)ij as T'ij, the components of stress T measured
in a frame which is deforming by Ft relative to the unprimed frame. That is to say, the basis vectors of the
primed frame are given by u'n = Ft un. In this case, since Ft is not a rotation, if the un start as unit
vectors, then the u'n are not unit vectors. One can think of each basis vector un as being aligned with its
own dumbbell dx(n) and then we have dx'(n)= Ft dx(n). What this says is that the basis vectors are
embedded in the fluid which deforms as it flows according to Ft. The basis vectors "convect" with the
fluid, so this upper Oldroyd stress is called the upper convective stress tensor.

For the lower Oldroyd stress in (K.5.19) we have

JL ≡ FtT T Ft (K.5.19) (K.5.30)

and this cannot be written in the form JL = R-1 M (R-1)T so this does not fit into our interpretive
template. But in the next section we show that JL is the covariant partner to the contravariant tensor JU so
they are both the same animal and we are happy to have the interpretation above for JU.

The Oldroyd convected stresses in developmental and standard notation

In the previous section two adjusted stress tensors were introduced,

JU ≡ Ft-1 T Ft-1T // upper

JL ≡ FtT T Ft . // lower (K.5.31)

In developmental notation, a covariant tensor gets an overbar while a contravariant one does not. If we
assume that frame S is Cartesian, then T = T̄ as was discussed for vectors in Section 5.9. We can interpret
the above two equations in this manner

404
 Appendix K: Deformation Tensors

J ≡ Ft-1 T Ft-1T // upper

J̄ ≡ FtT T̄ Ft // lower (K.5.32)

which we compare with (5.7.1) where we change generic matrix name M to T,

T' = R T RT // contravariant rank-2 tensor transforms this way

T̄' = ST T̄ S // covariant rank-2 tensor transforms this way . (K.5.33)

Setting R = Ft-1 and S = R-1 = Ft gives

T' = Ft-1 T Ft-1T // contravariant rank-2 tensor

T̄' = FtT T̄ Ft // covariant rank-2 tensor (K.5.34)

Therefore we identify

JU = J = T' = contravariant stress tensor T viewed from a frame convecting at Ft-1

JL = J̄ = T̄' = covariant stress tensor T viewed from a frame convecting at Ft-1 (K.5.35)

In Standard Notation the two equations

J ≡ Ft-1 T Ft-1T // upper

J̄ ≡ FtT T̄ Ft // lower (K.5.36)

become

Jij = (Ft-1)ia (Ft-1)jb Tab (JU)ij = Jij = contravariant

Jij = (Ft-1)ia (Ft-1)jb Tab (JL)ij = Jij = covariant (K.5.37)

and this explains the meaning of the words "upper" and "lower" in respect to the Oldroyd objects. See
footnote on Lai page 485.

405
 References

References

Listed in alphabetical order by last name of first author. A quick web search on article title can usually
locate documents with broken web links. Those below were last checked May 18, 2016.

L. A. Ahlfors, Complex Analysis, 2nd Ed. ( McGraw-Hill, New York, 1966).

G. Backus, Continuum Mechanics (Samizdat Press, Golden Colo., 1997).

J.D. Bjorken and S.D. Drell, Relativistic Quantum Mechanics (McGraw-Hill, New York, 1964).

E. B. Christoffel, "Ueber die Transformation der homogenen Differentialausdrücke zweiten Grades",

Journal für die reine and angewandte Mathemarik, 70 (1869), 46–70, 241–245. This paper may be found
in Christoffel's Collected Mathematical papers, Gesammelte Mathematische Abhandlungen, 2 vols.
(Tuebner, Leipzig-Berlin, 1910), downloadable from Google books.

R. Hermann, Ricci and Levi-Civita's Tensor Analysis Paper (English translation with comments) (Math
Sci Press, Brookline, MA, 1975, perhaps on-line) First, with much praise, Hermann provides an English
translation of the French Ricci & Levi-Civita paper referenced below, updating words, phrases and
symbols to the current day. Second, he inserts perhaps 180 pages of inline italicized text which translates
the ideas of the paper into mathematical frameworks not known or not connected to by the authors (eg,
fiber bundles, direct product spaces, Killing vectors, moving frames, group theory, etc.). Hermann
presents material that was more simply understood by later authors (eg, E. Cartan). Earlier in 1966
Hermann wrote a book Lie Groups for Physicists which contains the group theory chunk of this added
material. One realizes that differential geometry is a very large field touching upon many areas of
Mathematics (a house with many mansions).

W.M. Lai, D. Rubin and E. Krempl, Introduction to Continuum Mechanics, 4th Ed. (Butterworth-
Heinemann/Elsevier, Amsterdam, 2010). This book has had the same three authors since its first edition
in 1974, and it is with apologies to the last two authors that we reference the book just as Lai when we
should be saying Lai et.al. or perhaps LRK.

P. Lucht, Tensor Products, Wedge Products and Differential Forms (2016, http://user.xmission.com/~rimrock ).

H. Margenau and G.M. Murphy, The Mathematics of Physics and Chemistry, 2nd Ed. (D. van Nostrand,
London. 1956).

A. Messiah, Quantum Mechanics (John Wiley, New York, 1958). Reference is made to page 878 (Vol II)
of the North-Holland 1966 fifth printing paperback two-volume set, Chapter XX paragraph 2.

P. Moon and D.E. Spencer, Field Theory Handbook, Including Coordinate Systems, Differential
Equations and their Solutions (Springer-Verlag, Berlin, 1961). This is the place to find explicit
expressions for differential operators in many specific curvilinear coordinate systems. The book writes the
Laplace and Helmholtz equations in various systems and discusses separability and the special functions
that arise in the separated equations.

406
 References

P.M. Morse and H. Feshbach, Methods of Theoretical Physics ( McGraw-Hill, New York, 1953).

M.M.G. Ricci, T. Levi-Civita, "Méthodes de calcul différentiel absolu et leurs applications",

Mathematische Annalen (Springer) 54 (1–2): 125–201 (March 1900). This huge 77 page paper is
sometimes referred to as "the bible of tensor analysis". Levi-Civita was a student of Ricci and they
worked together on this paper and later elaborations. Their work was instrumental in Einstein's later
discovery of general relativity. The title is "Methods of absolute differential calculus and their
applications". Absolute differential calculus was the authors' phrase for what is now called Tensor
Analysis/Calculus/Algebra. The word absolute referred to the idea of equations being covariant (see
Section 7.15 above).

R. Sjamaar, Manifolds and Differential Forms (2015, http://www.math.cornell.edu/~sjamaar/manifolds ).

I. Stakgold, Boundary Value Problems of Mathematical Physics, Volumes 1 and 2 (Macmillan, London,
1967).

J.J. Sylvester, "On the General Theory of Associated Algebraical Forms" (Cambridge and Dublin Math.
Journal, VI, pp 289-293, 1851). This paper appears in H.F. Baker, Ed., The Collected Mathematical
Papers of James Joseph Sylvester (Cambridge University Press, 1901).

S. Weinberg, Gravitation and Cosmology: Principles and Applications of the General Theory of
Relativity (John Wiley & Sons, New York, 1972).

E. B. Wilson (notes of J.W. Gibbs), Vector Analysis (Dover, New York, 1960)

407