Notation, Concepts, and Conventions

in Relativity Theory

This chapter serves to briefly review the concepts relevant to the problems presented
in this book. Its purpose is to remind the reader of the basic concepts as well as
to introduce the notations and conventions that will be used. In particular, some
notations and conventions will vary throughout the different textbooks available on
the subject. Some of the different notations have been deliberately used in a number
of problems in order to familiarize the reader with the fact that different notations
occur in the literature.

General Notation
The components of a vector V will be written as V μ in contravariant form and
Vμ in covariant form with the index μ running over all the spacetime coordinates.
When referring to 3-vectors, Latin letters will be used for the spatial indices rather
than Greek ones, which we use for spacetime coordinates. In the case when an
explicit basis for a given vector space is needed, we will use the partial derivatives
∂μ to denote such a basis, i.e.,
V = V μ ∂μ . (0.1)
Similarly, tensor components will be denoted with superscripts for contravariant
indices and subscripts for covariant indices. Tensors with n indices, all down, are
called covariant tensors of rank n and tensors with n indices, all up, are called con-
travariant tensors of rank n. Tensors with indices both up and down are so-called
mixed tensors. Thus, a vector is a tensor of rank 1 (one index) and a scalar is a tensor
of rank zero (no indices). The Einstein summation convention is used throughout

1
2 Notation, Concepts, and Conventions in Relativity Theory

the book, implying that indices that are repeated are to be summed over the relevant
range. For example, in a four-dimensional spacetime, we have

3
V μ Uμ ≡ V μ Uμ = V 0 U0 + V 1 U1 + V 2 U2 + V 3 U3, (0.2)
μ=0

where U and V are vectors. Contravariant and covariant components are related by
lowering and raising with the metric tensor g = (gμν ) and its inverse g −1 = (g μν ),
respectively, i.e.,
Vμ = gμν V ν and V μ = g μν Vν . (0.3)
Partial derivatives of a given function f may be denoted in several ways, e.g.,
∂f
= ∂μ f = f,μ . (0.4)
∂x μ
Several indices after the comma in the latter notation represent higher-order deriva-
tives and the notation may also be used for vector components, for which indices
belonging to the vector component are written before the comma and indices denot-
ing derivatives after the comma, i.e.,
∂μ ∂ν f = f,μν and ∂μ Vν = Vν,μ . (0.5)
Objects with two indices may be represented in matrix form. We will indicate
this by putting parentheses around the considered objects. For example, we can
write the object A with two indices as
⎛ ⎞
A00 A01 A02 A03
⎜A10 A11 A12 A13 ⎟
A = (Aμν ) = ⎜ ⎝A20 A21 A22 A23 ⎠ .
⎟ (0.6)
A30 A31 A32 A33
By convention, the first index represents the row of the matrix and the second index
represents the column. When this is used for one covariant index and one con-
travariant index, the contravariant index is taken as the row index and the covariant
index as the column index.
For objects with more than two indices, we may use matrix notation to represent
parts of such objects by inserting a bullet (‘•’) in place of the indices being con-
μ μ
sidered. For example, the components A1ν of the three-index object Aλν would be
represented as the matrix
⎛ 0 ⎞
A10 A011 A012 A013
⎜A1 A1 A1 A1 ⎟
A•1• = (Aμ1ν ) = ⎜ 10 11
⎝A2 A2 A2 A2 ⎠ .
12 13 ⎟
(0.7)
10 11 12 13
A310 A311 A312 A313
 Special Relativity 3

Special Relativity
In a flat 1 + 3-dimensional spacetime and in Cartesian coordinates, the Minkowski
metric is given by
ds 2 = ημν dx μ dx ν = c2 dt 2 − dx 2 − dy 2 − dz2, (0.8)
where c is the speed of light in vacuum and x 0 = ct. In units of c = 1, so-called
natural units, it holds that x 0 = t. The metric tensor and its inverse, i.e., the inverse
metric tensor, can be written as
⎛ ⎞ ⎛ ⎞
1 0 0 0 1 0 0 0
⎜0 −1 0 0⎟ ⎜0 −1 0 0⎟
η = (ημν ) = ⎜ ⎟ −1
⎝0 0 −1 0 ⎠ ⇔ η = (η ) = ⎝0 0 −1 0 ⎠ .
μν ⎜ ⎟

0 0 0 −1 0 0 0 −1
(0.9)
For any two vectors x = (x μ ) = (x 0,x 1,x 2,x 3 ) and y = (y ν ) = (y 0,y 1,y 2,y 3 ) in
Minkowski space described by their contravariant components expressed in Carte-
sian coordinates, the Minkowski inner product is introduced as
x · y ≡ x 0 y 0 − x 1 y 1 − x 2 y 2 − x 3 y 3, (0.10)
which is obviously commutative, i.e., x · y = y · x. We also define the notation
x 2 ≡ x · x = (x 0 )2 − (x 1 )2 − (x 2 )2 − (x 3 )2 for the squared norm (‘length’) of
the vector x, which is indefinite, since it can be either positive or negative.1 The
Minkowski metric η and its inverse η−1 fulfill the relation
ημλ ηλν = ημν = ηνμ = δμν , (0.11)
where δμν is the Kronecker delta such that δμν = 1 if μ = ν and δμν = 0 if μ = ν.
We can write the Minkowski inner product in multiple ways as
x · y = x μ ημν y ν = ημν x μ y ν = x μ yμ = x 0 y0 + x i yi = x 0 y 0 − x i y i , (0.12)
where, e.g., x μ can be considered as the contravariant components of the vector x
and yμ the covariant components of the vector y, i.e., y0 = y 0 and yi = −y i , and
it also holds that x μ yμ = xν y ν . Furthermore, we say that the vector x is timelike if
x 2 > 0, lightlike if x 2 = 0, and spacelike if x 2 < 0. Note that lightlike vectors x
form a cone (x 0 )2 = (x 1 )2 + (x 2 )2 + (x 3 )2 and a nonspacelike vector x is future
pointing if x 0 > 0 and past pointing if x 0 < 0.

1 Note the abuse of notation – the symbol x 2 denotes both the ‘length’ of the vector x and the second spatial
contravariant component of the vector x. Unfortunately, this type of abuse of notation is difficult to avoid in
relativity theory, since the notation would otherwise be too cumbersome.
4 Notation, Concepts, and Conventions in Relativity Theory

In general, a Lorentz transformation  between two coordinate systems S and

S described by coordinates x and x  , respectively, is given by


x  = x ⇔ x μ = μν x ν . (0.13)

In particular, if the Lorentz transformation is a boost in the x 1 -direction, we can

write
⎛ ⎞ ⎛ ⎞
cosh θ − sinh θ 0 0 γ −βγ 0 0
⎜− sinh θ cosh θ 0 0⎟ ⎜−βγ γ 0 0⎟
(01) = ⎜⎝ 0
⎟=⎜
⎠ ⎝
⎟, (0.14)
0 1 0 0 0 1 0⎠
0 0 0 1 0 0 0 1
where θ is the rapidity, β ≡ v/c, and γ is the so-called gamma factor, i.e.,
γ ≡ γ (v) ≡ (1−v 2 /c2 )−1/2 , with v being the relative speed between the coordinate
systems S and S  . Furthermore, it holds that
1 v 1 v
cosh θ = γ = , sinh θ = βγ = , tanh θ = β = .
1− v 2 /c2 c 1− v 2 /c2 c
(0.15)
The formulas for (Lorentz) length contraction and time dilation are given by

 t
= = 1 − v 2 /c2, t = t  γ (v) = , (0.16)
γ (v) 1 − v 2 /c2
respectively.
The relativistic energy–momentum dispersion relation is given by
mc2
E  = mγ (v)c2 = , (0.17)
1 − v 2 /c2
where m is the mass of an object, v its speed, and E its energy. In the rest frame of
the object, it leads to Einstein’s famous formula

E = mc2 . (0.18)
The relativistic addition of velocities for colinear velocities v and v  is given by
v + v
v  = . (0.19)
1 + vv  /c2
In the nonrelativistic limit, i.e., v,v   c, the classical formula v   v + v  is
recovered.
Consider radiation of light in a specific coordinate direction of the coordinate
system S. One should think of the radiation as coming from a fixed source in
this coordinate system, where the radiation has frequency ν. For an observer in a
 Special Relativity 5

coordinate system S  moving along the same coordinate direction with the relative
velocity v, a frequency ν  is observed that is given by the relativistic Doppler
formula, i.e.,
c−v
ν = ν . (0.20)
c+v
If the observer is moving away from the source, there is a redshift in the frequency
of light, whereas if the observer is moving toward the source, there is a correspond-
ing blueshift.
For the (binary) reaction A + B −→ a + b + · · · , where two particles with
4-momenta pA and pB collide, using conservation of 4-momentum, we have
pA + pB = pa + pb + · · · , (0.21)
whereas for the decay A −→ a + b + · · · , we have the simpler relation
pA = pa + pb + pc + · · · . (0.22)
This can be generalized to any number of particles with corresponding 4-momenta
before and after a reaction, i.e.,
 
Pin = pi = pj = Pout . (0.23)
i=A,B,... j =a,b,...

Note that P 2 is invariant for any P = N k=1 pk , where N is the number of par-
ticles, and actually, for any two 4-vectors A and B, the Minkowski inner product
A · B = ημν Aμ B ν = Aμ Bμ is invariant under Lorentz transformations. Especially,
A2 = Aμ Aμ is invariant. This is useful in many applications.
In electromagnetism, the electromagnetic field strength tensor F is defined as
F μν = ∂ μ Aν − ∂ ν Aμ, (0.24)
where A = (Aμ ) = (φ,cA) is the 4-vector potential with φ and A = A(x) being
the electric scalar potential and the magnetic 3-vector potential, respectively, and
can be written as
⎛ ⎞
0 −E 1 −E 2 −E 3
⎜E 1 0 −cB 3 cB 2 ⎟
F = (F μν ) = ⎜
⎝E 2 cB 3
⎟, (0.25)
0 −cB 1 ⎠
E 3 −cB 2 cB 1 0
which is a real antisymmetric matrix, i.e., F μν = −F νμ , that combines both the
electric and magnetic field strengths, i.e., E = (E 1,E 2,E 3 ) and B = (B 1,B 2,B 3 ).
Using this tensor, Maxwell’s equations can be written as
∂μ F μν = j ν , ∂ μ F νλ + ∂ ν F λμ + ∂ λ F μν = 0, (0.26)
6 Notation, Concepts, and Conventions in Relativity Theory

where j = (j μ ) = (ρ,j) is the 4-current density with ρ = ρ(x) and j = j(x)

being the charge density and the electric 3-current density, respectively. In addition,
we have the two Lorentz invariants
F μν Fμν = 2 c2 B 2 − E 2 , μνλω F μν F λω = −8cB · E, (0.27)
where μνλω is the Levi-Civita tensor with  0123 = −0123 = 1. Maxwell’s equa-
tions describe how sources (charges and currents) give rise to electric and magnetic
fields. Assuming a moving test charge q with rest mass m and parametrizing the
trajectory of the test charge as x = x(s), where s is the proper time parameter,
the Lorentz force law describes how the field strengths determine the trajectory of
the test charge and is given by
mc2 ẍ μ (s) = q ẋν (s)F μν (x(s)), (0.28)
which is covariant under Lorentz transformations. The energy–momentum tensor
T of the electromagnetic field is defined as
0
T μν = 0 F μλ F λν + ημν Fλω F λω, (0.29)
4
where 0 is the electric constant (or permittivity of free space). It holds that
T is symmetric, i.e., T μν = T νμ , and T μμ = ημν T μν = 0. Furthermore, using
Maxwell’s equations, we obtain
∂μ T μν = 0 jμ F μν = −f ν , (0.30)
where f = (f μ ) = (j · E/c,ρE + j × B) is the Lorentz force density generated
by the 4-current j . Without (external) sources, i.e., when j = 0, T is conserved,
i.e., ∂μ T μν = 0.

General Relativity
In a curved spacetime, the metric is defined as
ds 2 = gμν dx μ dx ν , (0.31)
where the metric tensor and its inverse, i.e., the inverse metric tensor, are given by
g = (gμν ), g −1 = (g μν ), (0.32)
respectively. In the special case that the spacetime is flat (see Special Relativity),
we obtain
gμν = ημν , g μν = ημν , (0.33)
in Minkowski coordinates.
 General Relativity 7

The covariant derivatives of a covariant vector Aν and a contravariant vector Aν

are given by
∇μ Aν = Aν;μ = ∂μ Aν − μν
λ
Aλ, ∇μ Aν = Aν;μ = ∂μ Aν + μλ
ν
Aλ, (0.34)
respectively. In particular, it holds that ∇μ ∂ν = μν
λ λ
∂λ , where the coefficients μν
are called the Christoffel symbols of the second kind. Given a metric gμν = gνμ , the
Christoffel symbols of the Levi-Civita connection can be directly computed from
1
λ
μν = g λω ∂μ gνω + ∂ν gμω − ∂ω gμν . (0.35)
2
λ
In addition, it holds that μν = νμ
λ
, i.e., the Christoffel symbols are always
symmetric with respect to the two lower indices.
The parallel transport equation for a vector Aλ is given by
ẋ μ ∇μ Aλ = Ȧλ + μν ẋ A = 0,
λ μ ν
(0.36)
where the dot above (‘ ˙ ’) denotes differentiation with respect to the curve
parameter s. Furthermore, considering the Lagrangian density given by
L = gμν ẋ μ ẋ ν , (0.37)
and using the Euler–Lagrange equations, i.e.,
∂L d ∂L
μ
− = 0, (0.38)
∂x ds ∂ ẋ μ
where μ = 0,1, . . . ,n, we obtain the geodesic equations as
ẍ λ + μν ẋ ẋ = 0.
λ μ ν
(0.39)
Given three vector fields X, Y , and Z, the torsion T and the curvature R are
defined as
T (X,Y ) = ∇X Y − ∇Y X − [X,Y ] , (0.40)
R(X,Y )Z = [∇X,∇Y ] Z − ∇[X,Y ] Z. (0.41)
Both the torsion and curvature are tensors and are therefore linear in all of the
arguments X, Y , and Z, including when the arguments are multiplied by a scalar
function f , e.g.,
T (f X,Y ) = T (X,f Y ) = f T (X,Y ), T (X,Y + Z) = T (X,Y ) + T (X,Z).
(0.42)
Furthermore, the torsion and curvature tensors are antisymmetric in the arguments
X and Y as defined above
T (X,Y ) = −T (Y,X) and R(X,Y )Z = −R(Y,X)Z. (0.43)
8 Notation, Concepts, and Conventions in Relativity Theory

In local coordinates, we have

T (∂μ,∂ν )λ ∂λ = Tμν
λ
∂λ, R(∂μ,∂ν )∂λ = R ωλμν ∂ω, (0.44)

and the components of the torsion tensor and the Riemann curvature tensor may
be computed as
λ
Tμν = μν
λ
− νμ
λ
, (0.45)
ρ ω ρ
R ωλμν = ω
∂μ νλ − ω
∂ν μλ + ω
μρ νλ − νρ μλ, (0.46)
λ
respectively. Note that the Levi-Civita connection is torsion free as μν = νμ
λ
. For
fixed μ and ν, we can write the Riemann curvature tensor in matrix form as
 • 
R ••μν = ∂μ ν•
• •
− ∂ν μ• + μ• •
,ν• . (0.47)

Note that the Riemann curvature tensor is antisymmetric in μ and ν, i.e., R ••μν =
−R ••νμ , or in component form, R ωλμν = −R ωλνμ . If the torsion vanishes, i.e.,
T = 0, then we have the first Bianchi identity, i.e.,

R(X,Y )Z + R(Y,Z)X + R(Z,X)Y = 0, (0.48)

or in component form, we have

R ωλμν + R ωμνλ + R ωνλμ = 0. (0.49)

Furthermore, we have the second Bianchi identity in matrix form, i.e.,

 •   • •   • • 
∂μ R ••νλ + μ•,R ••νλ + ∂ν R ••λμ + ν•,R •λμ + ∂λ R ••μν + λ• ,R •μν = 0.
(0.50)
Using the Riemann curvature tensor, the Ricci tensor can be defined as

Rμν = R λμλν , (0.51)

which is symmetric, i.e., Rμν = Rνμ , and in turn, the Ricci scalar is defined as

R = g μν Rμν = Rμμ . (0.52)

Finally, the Einstein tensor is defined in terms of the Ricci tensor, the Ricci scalar,
and the metric tensor as
1
Gμν = Rμν − Rgμν . (0.53)
2
Note that it holds that the Einstein tensor is symmetric, i.e., Gμν = Gνμ , and con-
served, i.e., its covariant divergence vanishes ∇μ Gμν = 0. Under local coordinate
transformations y = y(x), we have
 General Relativity 9

 ∂x α ∂x β
gμν (y) = gαβ (x), (0.54)
∂y μ ∂y ν
∂x α ∂x β ∂y λ γ ∂y λ ∂ 2 x γ
  μν (y) =
λ
 (x) + , (0.55)
∂y μ ∂y ν ∂x γ αβ ∂x γ ∂y μ ∂y ν
∂y λ ∂x α ∂x β γ
T  μν (y) =
λ
T (x), (0.56)
∂x γ ∂y μ ∂y ν αβ
∂y ω ∂x γ ∂x α ∂x β δ
R
ω
λμν (y) = R (x). (0.57)
∂x δ ∂y λ ∂y μ ∂y ν γ αβ
Symmetries of a spacetime metric are associated to so-called Killing vector
fields. Consider a vector field X. By definition, X is a Killing vector field if
∇μ Xν + ∇ν Xμ = 0, (0.58)
for all indices μ and ν. Given a Killing vector field Xμ and a geodesic described by
coordinate functions x μ (s), the quantity
Q = ẋ μ Xμ = gμν ẋ μ Xν , (0.59)
is constant along the geodesic.
The dynamics of spacetime in vacuum are described in the Lagrange formalism
using the Einstein–Hilbert action, namely

MPl2
SEH = − R |ḡ| d 4 x, (0.60)
2
√
where MPl ≡ c2 / 8πG is the Planck mass, R is the Ricci scalar, and ḡ = det(g)
is the determinant of the metric tensor. For the case of a spacetime not in vacuum,
a matter contribution to the action is necessary

Smatter = L |ḡ| d 4 x, (0.61)

where L is the Lagrangian density of the matter contribution.

The Einstein gravitational field equations (or simply Einstein’s equations) follow
from the Euler–Lagrange equations for the action and are given by
8πG
Gμν =Tμν , (0.62)
c4
where G is Newton’s gravitational constant and T μν is the energy–momentum
tensor (or the stress–energy tensor) that describes the distribution of energy in
spacetime. The energy–momentum tensor is generally given by
2 δSmatter
Tμν = √ . (0.63)
|ḡ| δg μν
10 Notation, Concepts, and Conventions in Relativity Theory

For example, an external electromagnetic field gives a contribution to T μν such that

(see Special Relativity)
μν 0
TEM = 0 F μλ F λν + g μν Fλω F λω, (0.64)
4
whereas a perfect fluid (characterized by a 4-velocity u, a scalar density ρ0 , and a
scalar pressure p) gives
Tpfμν = (ρ0 + p)uμ uν − pg μν . (0.65)
In vacuum, Einstein’s equations reduce to Gμν = 0.
In the Newtonian limit and the weak field approximation, i.e., gμν  ημν +
hμν , where hμν is a small perturbation, the solutions to Einstein’s equations are
given by
2
h00 = h11 = h22 = h33 = , hμν = 0 ∀μ = ν, (0.66)
c2
where  is the gravitational potential for the matter distribution ρ and given
by  = −GM/r, that is the solution to the Newtonian equation ∇ 2  = 4πGρ.
Furthermore, the geodesic equations of motion become
d 2x i
= ∂ i  = −∂i . (0.67)
dt 2
The spherically symmetric vacuum solution to Einstein’s equations is the
Schwarzschild solution for which the Schwarzschild metric in spherical coordinates
is given by
   
2GM 2 2 2GM −1 2
ds = gμν dx dx = 1 − 2
2 μ ν
c dt − 1 − 2 dr − r 2 d2,
c r c r
(0.68)
where d2 describes the metric on a sphere, i.e., d2 = dθ 2 + sin2 θdφ 2 . For
large r, the Schwarzschild metric approaches the Minkowski metric. The partic-
ular value r = r∗ ≡ 2GM/c2 represents the Schwarzschild event horizon (or the
Schwarzschild radius) and is a coordinate singularity, i.e., it can be removed by
a change of coordinates. Such a coordinate change is given by Kruskal–Szekeres
coordinates u, v, θ , and φ, where θ and φ are the ordinary spherical coordinates on
a unit sphere S2 , the Kruskal–Szekeres metric is given by
16μ2 (2μ−r)/(2μ) 2GM
ds 2 = e dudv − r 2 d2, uv = (2μ − r)e(r−2μ)/(2μ) < ,
r c2 e
(0.69)
where μ ≡ GM/c2 .
 General Relativity 11

For a static spacetime, the metric can be written on the form

ds 2 = ϕ(x)2 dt 2 − gij (x)dx i dx j . (0.70)
Given two static observers A and B in this spacetime, signals sent from A to B with
frequencies fA and fB , respectively, will be redshifted according to
fA ϕ(xB )
z= −1= − 1. (0.71)
fB ϕ(xA )
In particular, in the Schwarzschild spacetime, a signal sent from a static observer at
r to an observer at infinity will be gravitationally redshifted according to
1 GM
z∞ ≡  −1 , (0.72)
1− 2GM c2 r
c2 r

where M is the mass of the gravitating body. More generally, the frequency f of a
light signal measured by an observer will be given by
f = gμν U μ N ν , (0.73)
where U is the 4-velocity of the observer and N the 4-frequency of the light signal,
which is parallel transported along the worldline of the light signal.
In cosmology, the cosmological principles are encoded into the Robertson–
Walker metric, which is given by
 
dr 2
ds = c dt − a(t)
2 2 2 2
+ r d ,
2 2
(0.74)
1 − kr 2
where a(t) is some function of the universal time t and k is a constant. By a suitable
coordinate transformation r → λr, one can always choose λ such that k takes one
of the three values k = 0, ±1. If k = 0, then the spatial part for any fixed t becomes
the Euclidean space R3 .
From Einstein’s equations, the assumption of the Robertson–Walker metric, and
the universe being filled by an ideal fluid, follow the two independent Friedmann
equations, namely
a  (t)2 + kc2 8πGρ + c2
= , (0.75)
a(t)2 3
 
a  (t) 4πG 3p c2
=− ρ+ 2 + , (0.76)
a(t) 3 c 3
where the first equation is derived from the 00-component of Einstein’s equations
and the second equation is derived from the first one together with the trace of
Einstein’s equations. Here  is the cosmological constant.
12 Notation, Concepts, and Conventions in Relativity Theory

Conventions
In this book, the following conventions will mainly be used in the presentation of
the problems and their corresponding solutions:

• Units: We will mostly use units in which the speed of light in vacuum c has been
set to c = 1, these are usually known as natural units. In some problems, we have
also set h̄ = 1, if relevant. Normally, we do not use units in which Newton’s
gravitational constant G has been set to G = 1. In some problems, it is useful to
use SI units.
• Vectors: In a four-dimensional spacetime, we will normally denote a 4-vector A
by its contravariant components as follows

A = (Aμ ) = (A0,A1,A2,A3 ), (0.77)

where A0 is the temporal component and Ai (i = 1,2,3) are the spatial compo-
nents, which is related to the 4-vector expressed in its covariant components as
follows

(Aμ ) = (A0,A1,A2,A3 ), (0.78)

where it holds that Aμ = gμν Aν with (gμν ) being the given metric tensor. In some
textbooks, the convention that the temporal component of a 4-vector is written
as the last component of the vector is used, i.e., A = (A1,A2,A3,A4 ), whereas
in other textbooks, the convention that the standard components of a 4-vector
are chosen as its covariant components might be used. We will not use these
conventions.
• Metrics: In four-dimensional spacetimes, we adopt the convention that the sig-
nature is + − −− and, when relevant, place the temporal direction first and
denote it by 0. Therefore, in standard coordinates on Minkowski space, the metric
tensor is (ημν ) = diag(1,−1,−1,−1) (see Special Relativity) and its inverse is
given by (ημν ) = diag(1,−1,−1,−1). Thus, we have Aμ = ημν Aν , where
A0 = η00 A0 = A0 and Ai = ηii Ai = −Ai (for fixed i = 1,2,3). In a gen-
eral coordinates, the metric is given by ds 2 = gμν dx μ dx ν (see General Relativ-
ity) and the metric tensor components represented in matrix form as g = (gμν )
from which its corresponding inverse components g −1 = (g μν ) can be computed.
It must hold that gg −1 = g −1 g = 14 , where 14 is the 4 × 4 identity matrix.
• Sign convention of the Levi-Civita pseudotensor: We define  0123 = +1 (see
Special Relativity), which means that with our convention for the Minkowski
metric, we have 0123 = −1.
• Partial derivatives: We will mostly denote covariant and contravariant partial
derivatives as
 Conventions 13
∂
∂μ ≡ , ∂ μ ≡ g μν ∂ν , (0.79)
∂x μ
where (g μν ) is the inverse metric tensor.
• Covariant derivatives and Christoffel symbols: For covariant derivatives (see
General Relativity), we will mostly use the notation ∇μ , but the notation Dμ will
sometimes be used. For Christoffel symbols of the second kind, we will only use
λ
the notation μν (see General Relativity).
• Sign convention of the Ricci tensor: The Ricci tensor may sometimes be defined
as Rμν = R λ μνλ , which introduces a sign difference to our definition (see General
Relativity) due to the antisymmetry of the Riemann curvature tensor as
R λ μνλ = −R λ μλν . (0.80)
• Sign convention of Einstein’s equations: There is a sign convention in the
definition of Einstein’s equations, i.e., Gμν = ±8πGTμν /c4 , where we use the
positive sign.
In general, it is important to keep in mind that different texts may use different
conventions. In particular, the sign discrepancies in different expressions will often
be due to differing sign conventions of the metric, the Levi-Civita pseudotensor, the
Ricci tensor, and Einstein’s equations.