This action might not be possible to undo. Are you sure you want to continue?

General

Relativity

Version 3.2

Alan Macdonald

Luther College, Decorah, IA USA

mailto:macdonal@luther.edu

http://faculty.luther.edu/

∼

macdonal

c (

To Ellen

“The magic of this theory will hardly fail to impose itself on anybody

who has truly understood it.”

Albert Einstein, 1915

“The foundation of general relativity appeared to me then [1915],

and it still does, the greatest feat of human thinking about Nature,

the most amazing combination of philosophical penetration, physical

intuition, and mathematical skill.”

Max Born, 1955

“One of the principal objects of theoretical research in any depart-

ment of knowledge is to ﬁnd the point of view from which the subject

appears in its greatest simplicity.”

Josiah Willard Gibbs

“There is a widespread indiﬀerence to attempts to put accepted the-

ories on better logical foundations and to clarify their experimental

basis, an indiﬀerence occasionally amounting to hostility. I am con-

cerned with the eﬀects of our neglect of foundations on the educa-

tion of scientists. It is plain that the clearer the teacher, the more

transparent his logic, the fewer and more decisive the number of ex-

periments to be examined in detail, the faster will the pupil learn

and the surer and sounder will be his grasp of the subject.”

Sir Hermann Bondi

“Things should be made as simple as possible, but not simpler.”

Albert Einstein

Preface

My purpose here is to provide, with a minimum of mathematical machinery

and in the fewest possible pages, a clear and careful explanation of the physical

principles and applications of classical general relativity. The prerequisites are

single variable calculus, a few basic facts about partial derivatives and line

integrals, a little matrix algebra, and some basic physics.

The book is for those seeking a conceptual understanding of the theory, not

computational prowess. Despite it’s brevity and modest prerequisites, it is a

serious introduction to the physics and mathematics of general relativity which

demands careful study. The book can stand alone as an introduction to general

relativity or it can be used as an adjunct to standard texts.

Chapter 1 is a self-contained introduction to those parts of special relativ-

ity we require for general relativity. The approach was chosen to facilitate the

transition to general relativity in Chapter 2. The ﬁrst two chapters systemati-

cally exploit the mathematical analogy which led to general relativity: a curved

spacetime is to a ﬂat spacetime as a curved surface is to a ﬂat surface. Before

introducing a spacetime concept, its analog for surfaces is presented. This is

not a new idea, but it is used here more systematically than elsewhere. For

example, when the metric ds of general relativity is introduced in Chapter 2,

the reader has already seen a metric in three other contexts.

Chapter 2 introduces the physical principles on which general relativity is

based. The basic concepts of Riemannian geometry are developed in order

to express these principles mathematically as postulates. The purpose of the

postulates is not to achieve complete rigor – which is neither desirable nor

possible in a book at this level – but to state clearly the physical principles,

and to exhibit clearly the relationship to special relativity and the analogy with

surfaces. The postulates are in one-to-one correspondence with the fundamental

concepts of Riemannian geometry: manifold, metric, geodesic, and curvature.

We take a nonstandard approach to the metric of relativity, which is similar

to the standard approach to the metric in Euclidean geometry. In geometry,

distance is ﬁrst understood geometrically, independently of any coordinate sys-

tem. Then the Pythagorean theorem is proved. The theorem is important, but

the geometric notion of distance is fundamental. Similarly, we deﬁne the met-

ric of spacial and general relativity physically, independently of any coordinate

system. This puts the physical meaning of the metric front and center, where it

belongs. We then prove that there is a metric g

jk

so that ds

2

= g

jk

(y

i

) dy

j

dy

k

.

I believe that this approach provides easier access to and deeper understanding

of relativity, and greatly simpliﬁes the development of the theory. In particu-

lar, tensors are not used. (Similarly, modern elementary diﬀerential geometry

texts often develop the intrinsic geometry of curved surfaces by focusing on the

geometric meaning of the metric. Tensors are not used.)

Chapter 3 solves the ﬁeld equation for a spherically symmetric spacetime

to obtain the Schwartzschild metric. The geodesic equations are then solved

and applied to the classical solar system tests of general relativity. There is a

discussion of the Kerr metric, including gravitomagnetism and its observation

by the LAGEOS satellites. The chapter closes with short sections on the binary

pulsar and black holes. In this chapter, as elsewhere, I have tried to provide the

cleanest possible calculations.

Chapter 4 applies general relativity to cosmology. We obtain the Robertson-

Walker metric in an elementary manner without using the ﬁeld equation. We

then solve the ﬁeld equation with a nonzero cosmological constant for a ﬂat

Robertson-Walker spacetime. WMAP data allows us to determine all unknown

parameters in the solution, giving the new “standard model” of the universe

with dark matter and dark energy.

There have been many spectacular astronomical discoveries and observa-

tions since 1960 which are relevant to general relativity. We describe them at

appropriate places in the book.

Some 50 exercises are scattered throughout. They often serve as examples

of concepts introduced in the text. If they are not done, they should be read.

Some tedious (but always straightforward) calculations have been omitted.

They are best carried out with a computer algebra system. Some material

has been placed in 14 appendices to keep the main line of development visible.

The appendices occasionally require more background of the reader than the

text, but they may be omitted without loss of continuity. Appendix 1 gives the

values of various physical constants. Appendix 2 contains several approximation

formulas used in the text.

Contents

Preface

Contents

1 Flat Spacetimes

1.1 Spacetimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

1.2 The Inertial Frame Postulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

1.3 The Metric Postulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

1.4 The Geodesic Postulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

2 Curved Spacetimes

2.1 History of Theories of Gravity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28

2.2 The Key to General Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

2.3 The Local Inertial Frame Postulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

2.4 The Metric Postulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

2.5 The Geodesic Postulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

2.6 The Field Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

3 Spherically Symmetric Spacetimes

3.1 Stellar Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

3.2 The Schwartzschild Metric . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

3.3 The Solar System Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

3.4 Kerr Spacetimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61

3.5 The Binary Pulsar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

3.6 Black Holes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

4 Cosmological Spacetimes

4.1 Our Universe I . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

4.2 Robertson-Walker Spacetimes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

4.3 The Expansion Redshift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

4.4 Our Universe II . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

4.5 General Relativity Today . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80

Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

Index

Chapter 1

Flat Spacetimes

1.1 Spacetimes

The general theory of relativity is our best theory of space, time, and gravity. It

is commonly felt to be the most beautiful of all physical theories. Albert Einstein

created the theory during the decade following the publication, in 1905, of his

special theory of relativity. The special theory is a theory of space and time

which applies when gravity is insigniﬁcant. The general theory generalizes the

special theory to include gravity.

In geometry the fundamental entities are points. A point is a speciﬁc place.

In relativity the fundamental entities are events. An event is a speciﬁc time and

place. For example, the collision of two particles is an event. A concert is an

event (idealizing it to a single time and place). To attend the concert, you must

be at the time and the place of the event.

A ﬂat or curved surface is a set of points. (We shall prefer the term “ﬂat

surface” to “plane”.) Similarly, a spacetime is a set of events. For example,

we might consider the events in a speciﬁc room between two speciﬁc times. A

ﬂat spacetime is one without signiﬁcant gravity. Special relativity describes

ﬂat spacetimes. A curved spacetime is one with signiﬁcant gravity. General

relativity describes curved spacetimes.

There is nothing mysterious about the words “ﬂat” or “curved” attached

to a set of events. They are chosen because of a remarkable analogy, already

hinted at, concerning the mathematical description of a spacetime: a curved

spacetime is to a ﬂat spacetime as a curved surface is to a ﬂat surface. This

analogy will be a major theme of this book; we shall use the mathematics of

ﬂat and curved surfaces to guide our understanding of the mathematics of ﬂat

and curved spacetimes.

We shall explore spacetimes with clocks to measure time and rods (rulers)

to measure space, i.e., distance. However, clocks and rods do not in fact live

up to all that we usually expect of them. In this section we shall see what we

expect of them in relativity.

11

1.1 Spacetimes

Clocks. First a deﬁnition. A curve is a continuous succession of points

in a surface. Similarly, a worldline is a continuous succession of events in a

spacetime. A moving particle or a pulse of light emitted in a single direction is

present at a continuous succession of events, its worldline. Even if a particle is

at rest, time passes, and the particle has a worldline.

The length of a curve between two given points depends on the curve. Sim-

ilarly, the time between two given events measured by a clock moving between

the events depends on the clock’s worldline! J. C. Hafele and R. Keating pro-

vided the most direct veriﬁcation of this in 1971. They brought two atomic

clocks together, placed them in separate airplanes which circled the Earth in

opposite directions, and brought the clocks together again. Both clocks moved

between the event of their separation and the event of their reunion. The clocks

measured diﬀerent times between the events. The diﬀerence was small, about

10

−7

sec, but was well within the ability of the clocks to measure. There is no

doubt that the eﬀect is real.

Relativity predicts the measured diﬀerence. Exercise 1.9 shows that special

relativity predicts a diﬀerence between the clocks. Exercise 2.1 shows that

general relativity predicts a further diﬀerence. Exercise 3.3 shows that general

relativity predicts the observed diﬀerence. Relativity prtedicts large diﬀerences

between clocks whose relative velocity is close to that of light.

The best answer to the question “How can the clocks in the experiment

possibly disagree?” is the question “Why should they agree?” After all, the

clocks are not connected. According to everyday ideas they should agree because

there is a universal time, common to all observers. It is the duty of clocks to

report this time. The concept of a universal time was abstracted from experience

with small speeds (compared to that of light) and clocks of ordinary accuracy,

where it is nearly valid. The concept permeates our daily lives; there are clocks

everywhere telling us the time. However, the Hafele-Keating experiment shows

that there is no universal time. Time, like distance, is route dependent.

Since clocks on diﬀerent worldlines between two events measure diﬀerent

times between the events, we cannot speak of the time between two events.

However, the relative rates of processes - the ticking of a clock, the frequency of

a tuning fork, the aging of an organism, etc. - are the same along all worldlines.

(Unless some adverse physical condition aﬀects a rate.) Twins traveling in the

two airplanes of the Hafele-Keating experiment would each age according to

the clock in their airplane. They would thus be of slightly diﬀerent ages when

reunited.

12

1.1 Spacetimes

Rods. We need another deﬁnition. Consider astronauts in interstellar space,

where gravity is insigniﬁcant. If their rocket is not ﬁring and their ship is not

spinning, then they will feel no forces acting on them and they can ﬂoat freely in

their cabin. If their spaceship is accelerating, then they will feel a force pushing

them back against their seat. If the ship turns to the left, then they will feel

a force to the right. If the ship is spinning, they will feel a force outward from

the axis of spin. Call these forces inertial forces.

Fig. 1.1: An accelerom-

eter. The weight is

held at the center by

springs. Acceleration

causes the weight to

move from the center.

Accelerometers measure inertial forces. Fig. 1.1

shows a simple accelerometer that detects any motion

of a weight held at the center of a frame by springs.

An inertial object experiences no inertial forces. If an

object is inertial, then any object moving at a constant

velocity with respect to it is also inertial.

In special relativity we make an assumption which

allows us to speak of the distance between two inertial

objects at rest with respect to each other: Inertial rigid

rods side by side and at rest with respect to two inertial

objects measure the same distance between the objects.

More precisely, we assume that any diﬀerence is due to

some adverse physical cause (e.g., thermal expansion)

to which an “ideal” rigid rod would not be subject. In

particular, the history of a rigid rod does not aﬀect its

length. Noninertial rods are diﬃcult to deal with in

relativity, and we shall not consider them.

In the next three sections we give three postulates for special relativity.

The inertial frame postulate asserts that certain natural coordinate systems,

called inertial frames, exist for a ﬂat spacetime. The metric postulate asserts

a universal light speed and a slowing of clocks moving in inertial frames. The

geodesic postulate asserts that inertial particles and light move in a straight line

at constant speed in inertial frames.

We shall use the analogy mentioned above to help us understand the pos-

tulates. Imagine two dimensional beings living in a ﬂat surface. These surface

dwellers can no more imagine leaving their two spatial dimensions than we can

imagine leaving our three spatial dimensions. Before introducing a postulate

for a ﬂat spacetime, we introduce the analogous postulate formulated by sur-

face dwellers for a ﬂat surface. The postulates for a ﬂat spacetime use a time

dimension, but those for a ﬂat surface do not.

13

1.2 The Inertial Frame Postulate

1.2 The Inertial Frame Postulate

Surface dwellers ﬁnd it useful to label the points of their ﬂat surface with co-

ordinates. They construct, using identical rigid rods, a square grid and assign

rectangular coordinates (x, y) to the nodes of the grid in the usual way. See

Fig. 1.2. They specify a point by using the coordinates of the node nearest

the point. If more accurate coordinates are required, they build a ﬁner grid.

Surface dwellers call the coordinate system a planar frame. They postulate:

The planar frame postulate for a ﬂat surface

A planar frame can be constructed with any point P as origin and

with any orientation.

Fig. 1.2: A planar frame.

Similarly, it is useful to label the events in a ﬂat

spacetime with coordinates (t, x, y, z). The coordi-

nates specify when and where the event occurs, i.e.,

they completely specify the event. We now describe

how to attach coordinates to events. The procedure

is idealized, but it gives a clear physical meaning to

the coordinates.

To specify where an event occurs, construct, us-

ing identical rigid rods, an inertial cubical lattice.

See Fig. 1.3. Assign rectangular coordinates (x, y, z)

to the nodes of the lattice in the usual way. Specify

where an event occurs by using the coordinates of

the node nearest the event.

Fig. 1.3: An inertial lat-

tice.

To specify when an event occurs, place a clock at

each node of the lattice. Then the times of events at

a given node can be speciﬁed by reading the clock at

that node. But to compare meaningfully the times of

events at diﬀerent nodes, the clocks must be in some

sense synchronized. As we shall see soon, this is not

a trivial matter. (Remember, there is no universal

time.) For now, assume that the clocks have been

synchronized. Then specify when an event occurs by

using the time, t, on the clock at the node nearest the

event. And measure the coordinate time diﬀerence

∆t between two events using the synchronized clocks

at the nodes where the events occur. Note that this

requires two clocks.

The four dimensional coordinate system obtained in this way from an inertial

cubical lattice with synchronized clocks is called an inertial frame. The event

(t, x, y, z) = (0, 0, 0, 0) is the origin of the inertial frame. We postulate:

The inertial frame postulate for a ﬂat spacetime

An inertial frame can be constructed with any event E as origin,

with any orientation, and with any inertial object at E at rest in it.

14

1.2 The Inertial Frame Postulate

If we suppress one or two of the spatial coordinates of an inertial frame, then

we can draw a spacetime diagram and depict worldlines. For example, Fig. 1.4

shows the worldlines of two particles. One is at rest on the x-axis and one moves

away from x = 0 and then returns more slowly.

Fig. 1.4: Two worldlines. Fig. 1.5: Worldline of a particle moving

with constant speed.

Exercise 1.1. Show that the worldline of an object moving with constant

speed v along the x-axis is a straight line with slope v. See Fig. 1.5.

We return to the matter of synchronizing the clocks in the lattice. First of

all, what does it mean to say that separated clocks are synchronized? It was

Einstein who ﬁrst realized that the answer to this question is not given to us by

Nature; rather, it must be answered with a deﬁnition.

Exercise 1.2. Why not simply bring the clocks together, synchronize them,

move them to the nodes of the lattice, and call them synchronized?

We might try the following deﬁnition. Send a signal from a node P of the

lattice at time t

P

according to the clock at P. Let it arrive at a node Q of

the lattice at time t

Q

according to the clock at Q. Let the distance between

the nodes be D and the speed of the signal be v. Say that the clocks are

synchronized if

t

Q

= t

P

+D/v. (1.1)

Intuitively, the term D/v compensates for the time it takes the signal to get to

Q. This deﬁnition is ﬂawed because it contains a logical circle: v is deﬁned by

a rearrangement of Eq. (1.1): v = D/(t

Q

−t

P

). Synchronized clocks cannot be

deﬁned using v because synchronized clocks are needed to deﬁne v.

We adopt the following deﬁnition, essentially due to Einstein. Emit a pulse

of light from a node P at time t

P

according to the clock at P. Let it arrive at a

node Q at time t

Q

according to the clock at Q. Similarly, emit a pulse of light

from Q at time t

Q

and let it arrive at P at t

P

. The clocks are synchronized if

t

Q

−t

P

= t

P

−t

Q

, (1.2)

i.e., if the times in the two directions are equal.

A reformulation of the deﬁnition will make it more transparent. Suppose

the pulse from Q to P is the reﬂection of the pulse from P to Q. Then t

Q

= t

Q

15

1.2 The Inertial Frame Postulate

in Eq. (1.2). Let 2T be the round trip time: 2T = t

P

− t

P

. Substitute this in

Eq. (1.2):

t

Q

= t

P

+T; (1.3)

the clocks are synchronized if the pulse arrives at Q in half the time it takes for

the round trip.

Exercise 1.3. Explain why Eq. (1.3) is a satisfactory deﬁnition but Eq.

(1.1) is not.

There is a tacit assumption in the deﬁnition of synchronized clocks that the

two sides of Eq. (1.2) do not depend on the times that the pulses are sent:

Emit pulses of light from a node R at times t

R

and t

R

according

to a clock at R. Let them arrive at a node S at times t

S

and t

S

according to a clock at S. Then

t

S

−t

R

= t

S

−t

R

. (1.4)

With this assumption we can be sure that synchronized clocks will remain syn-

chronized.

Exercise 1.4. Show that with the assumption Eq. (1.4), T in Eq. (1.3) is

independent of the time the pulse is sent.

A rearrangement of Eq. (1.4) gives

∆s

o

= ∆s

e

, (1.5)

where ∆s

o

= t

S

−t

S

is the time between the observation of the pulses at S and

∆s

e

= t

R

−t

R

is the time between the emission of the pulses at R. (We use ∆s

rather than ∆t to conform to notation used later in more general situations.) If

a clock at R emits pulses of light at regular intervals to S, then Eq. (1.5) states

that an observer at S sees (actually sees) the clock at R going at the same rate

as his clock. Of course, the observer at S will see all physical processes at R

proceed at the same rate they do at S.

We will encounter situations in which ∆s

o

= ∆s

e

. Deﬁne the redshift

z =

∆s

o

∆s

e

−1. (1.6)

Equations (1.4) and (1.5) correspond to z = 0. If, for example, z = 1 (∆s

o

/∆s

e

= 2), then the observer at S would see clocks at R, and all other physical

processes at R, proceed at half the rate they do at S.

If the two “pulses” of light in Eq. (1.6) are successive wavecrests of light

emitted at frequency f

e

= (∆s

e

)

−1

and observed at frequency f

o

= (∆s

o

)

−1

,

then Eq. (1.6) can be written

z =

f

e

f

o

−1. (1.7)

16

1.2 The Inertial Frame Postulate

In Exercise 1.6 we shall see that Eq. (1.5) is violated, i.e., z = 0, if the

emitter and observer are in relative motion in a ﬂat spacetime. This is called

a Doppler redshift. Later we shall see two other kinds of redshift: gravitational

redshifts in Sec. 2.2 and expansion redshifts in Sec. 4.1. The three types of

redshifts have diﬀerent physical origins and so must be carefully distinguished.

The inertial frame postulate asserts in part

Fig. 1.6: Light traversing a tri-

angle in opposite directions.

that clocks in an inertial lattice can be synchro-

nized according to the deﬁnition Eq. (1.2), or,

in P. W. Bridgeman’s descriptive phrase, we

can “spread time over space”. We now prove

this with the aid of an auxiliary assumption.

The reader may skip the proof and turn to the

next section without loss of continuity.

Let 2T be the time, as measured by a clock

at the origin O of the lattice, for light to travel

from O to another node Q and return after

being reﬂected at Q. Emit a pulse of light at O

toward Q at time t

O

according to the clock at

O. When the pulse arrives at Q set the clock

there to t

Q

= t

O

+ T. According to Eq. (1.3)

the clocks at O and Q are now synchronized.

Synchronize all clocks with the one at O in this way.

To show that the clocks at any two nodes P and Q are now synchronized

with each other, we must make this assumption:

The time it takes light to traverse a triangle in the lattice is inde-

pendent of the direction taken around the triangle.

See Fig. 1.6. In an experiment performed in 1963, W. M. Macek and D. T. M.

Davis, Jr. veriﬁed the assumption for a square to one part in 10

12

. See 3.

Reﬂect a pulse of light around the triangle OPQ. Let the pulse be at

O, P, Q, O at times t

O

, t

P

, t

Q

, t

R

according to the clock at that node. Similarly,

let the times for a pulse sent around in the other direction be t

O

, t

Q

, t

P

, t

R

.

See Fig. 1.6. We have the algebraic identities

t

R

−t

O

= (t

R

−t

P

) + (t

P

−t

Q

) + (t

Q

−t

O

)

t

R

−t

O

= (t

R

−t

P

) + (t

P

−t

Q

) + (t

Q

−t

O

). (1.8)

According to our assumption, the left sides of the two equations are equal. Also,

since the clock at O is synchronized with those at P and Q,

t

P

−t

O

= t

R

−t

P

and t

R

−t

Q

= t

Q

−t

O

.

Thus, subtracting the equations Eq. (1.8) gives

t

Q

−t

P

= t

P

−t

Q

,

i.e., the clocks at P and Q are synchronized.

17

1.3 The Metric Postulate

1.3 The Metric Postulate

Let P and Q be points in a ﬂat surface. Diﬀerent curves between the points have

diﬀerent lengths. But surface dwellers single out for special study the distance

∆s along the straight line between P and Q. They call ∆s the proper distance

between the points.

The proper distance ∆s is deﬁned geometrically, independently of any planar

frame. But surface dwellers discover a simple formula for ∆s in terms of the

coordinate diﬀerences of two points in a planar frame:

The metric postulate for a ﬂat surface

Let ∆s be the proper distance between points P and Q. Let P and

Q have coordinate diﬀerences (∆x, ∆y) in a planar frame. Then

∆s

2

= ∆x

2

+ ∆y

2

. (1.9)

Fig. 1.7: ∆s is given by Eq. (1.9)

in both planar frames.

The coordinate diﬀerences ∆x and ∆y be-

tween P and Q are diﬀerent in diﬀerent planar

frames. See Fig. 1.7. However, the particular

combination of the diﬀerences in Eq. (1.9) will

always produce ∆s. Neither ∆x nor ∆y has a

geometric signiﬁcance independent of the par-

ticular planar frame chosen. The two of them

together do: they determine ∆s, which has a di-

rect geometric signiﬁcance, independent of any

coordinate system.

Given two events in a ﬂat spacetime, there

is a quantity ∆s relating them. It is called the

(spacetime) interval between the events. The

deﬁnition of ∆s in a ﬂat spacetime is more complicated than in a ﬂat surface,

as there are three ways in which events E and F can be, we say, separated:

• Lightlike separated. E and F can be on the worldline of a pulse of light.

Set ∆s = 0 for E and F. This peculiar deﬁnition will bring uniﬁcation to

the three kinds of separation.

• Timelike separated. E and F can be on the worldline of an inertial clock.

Let ∆s be the time between E and F measured by an inertial clock. This

is the proper time between the events. Other clocks moving between the

events will measure diﬀerent times. But we single out for special study

the proper time ∆s .

• Spacelike separated. An inertial rigid rod can have its ends present at E

and F simultaneously. (Simultaneously means that light pulses emitted

at E and F reach the center of the rod simultaneously. Equivalently, E

and F are simultaneous in an inertial frame in which the rod is at rest.)

Let [∆s [ be the length the rod. The reason for the absolute value will

become clear later. This is the proper distance between the events.

18

1.3 The Metric Postulate

The spacetime interval ∆s is deﬁned physically, independently of any inertial

frame. But there is a simple formula for ∆s in terms of the coordinate diﬀerences

of two events in an inertial frame:

The metric postulate for a ﬂat spacetime

Let ∆s be the interval between events E and F. Let the events have

coordinate diﬀerences (∆t, ∆x, ∆y, ∆z) in an inertial frame. Then

∆s

2

= ∆t

2

−∆x

2

−∆y

2

−∆z

2

. (1.10)

The coordinate diﬀerences between E and F, including the time coordinate

diﬀerence, are diﬀerent in diﬀerent inertial frames. For example, suppose that

an inertial clock measures a proper time ∆s between two events. In an inertial

frame in which the clock is at rest, ∆t = ∆s and ∆x = ∆y = ∆z = 0. This will

not be the case in an inertial frame in which the clock is moving. However, the

particular combination of the diﬀerences in Eq. (1.10) will always produce ∆s.

No one of the coordinate diﬀerences has a physical signiﬁcance independent of

the particular inertial frame chosen. The four of them together do: they deter-

mine ∆s, which has a direct physical signiﬁcance, independent of any inertial

frame.

This shows that the joining of space and time into spacetime is not an

artiﬁcial technical trick. Rather, in the words of Hermann Minkowski, who

introduced the spacetime concept in 1908, “Space by itself, and time by itself,

are doomed to fade away into mere shadows, and only a kind of union of the

two will preserve an independent reality.”

The metric postulate is based on the fundamental assumption of a universal

light speed: the speed of light has always the same value c in all inertial frames.

With a suitable choice of units of time and space we can choose c = 1 . Choose

the second as the unit of time. Then choose the distance light travels in one

second – about 310

10

cm – as the unit of distance. Call this one (light) second

of distance. (You are probably familiar with a similar unit of distance - the light

year.) Then 1 cm = 3.3 10

−11

sec. With this convention c = 1, and all other

speeds are expressed as a fraction of the speed of light. Ordinarily the fractions

are very small.

We show that a universal light speed implies the metric postulate, Eq. (1.11).

It will be convenient to drop the y- and z-coordiantes in Eq. (1.10) and use

∆s

2

= ∆t

2

−∆x

2

. (1.11)

Lightlike separated events. In this case E and F can be on the worldline

of a pulse of light. With c = 1, a universal light speed means c = [∆x/∆t[ = 1 ,

i.e., [∆x[ = [∆t[. Since by deﬁnition ∆s = 0 for lightlike separated events, Eq.

(1.11) is satisﬁed.

19

1.3 The Metric Postulate

Timelike separated events. In this case E and F can be on the worldline

of an inertial clock C. By deﬁnition, ∆s is the time C measures between the

events. We show that Eq. (1.11) is satisﬁed.

Fig. 1.8: ∆s

2

= ∆t

2

−∆x

2

for timelike separated

events. e and f are the points at which events E

and F occur.

Let C carry a rod R perpen-

dicular to its direction of mo-

tion. Let R have a mirror M on

its end. At E a light pulse is sent

along R from C. The length of

R is chosen so that the pulse is

reﬂected by M back to F. Fig.

1.8 shows the path of the light in

I, together with C, R, and M as

the light reﬂects oﬀ M.

Refer to the rightmost trian-

gle in Fig. 1.8. In I, the distance

between E and F is ∆x. This gives the labeling of the base of the triangle. In

I, the light takes the time ∆t from E to M to F. Since c = 1 in I, the light

travels a distance ∆t in I. This gives the labeling of the hypotenuse. C is at

rest in some inertial frame I

. In I

**, the light travels the length of the rod twice
**

in the proper time ∆s between E and F measured by C. Since c = 1 in I

,

the length of the rod is

1

2

∆s in I

**. This gives the labeling of the altitude of the
**

triangle. (There is a tacit assumption here that the length of R is the same in I

and I

**. Appendix 4 discusses this.) Applying the Pythagorean theorem to the
**

triangle shows that Eq. (1.11) is satisﬁed for timelike separated events.

In short, since the light travels farther in I than in I

**(the hypotenuse twice
**

vs. the altitude twice) and the speed c = 1 is the same in I and I

, the time (=

distance/speed) between E and F is longer in I than for C. This shows, in a

most graphic way, that accepting a universal light speed forces us to abandon a

universal time.

The argument shows how it is possible for a single pulse of light to have

the same speed in inertial frames moving with respect to each other: the speed

(distance/time) can be the same because the distance and the time are diﬀerent

in the two frames.

Exercise 1.5. Criticize the following argument. We have just seen that

the time between two events is greater in I than in I

**. But exactly the same
**

argument carried out in I

**will show that the time between the events is greater
**

in I

**than in I. This is a contradiction.
**

We can express the proper time ∆s in a diﬀerent way. Let v = [∆x[/∆t be

the speed of an inertial clock. Then from Eq. (1.11), the proper time is

∆s = (∆t

2

−∆x

2

)

1

2

= [1 −(∆x/∆t)

2

]

1

2

∆t = (1 −v

2

)

1

2

∆t. (1.12)

The proper time between two events is less than the time determined by the

synchronized clocks of an inertial frame: ∆s < ∆t. Informally, “moving clocks

run slowly”. This is called time dilation. According to Eq. (1.12) the time

20

1.3 The Metric Postulate

dilation factor is (1 − v

2

)

1

2

. For normal speeds, v is very small, v

2

is even

smaller, and so from Eq. (1.12), ∆s ≈ ∆t, as expected. But as v →1, ∆t/∆s =

(1 −v

2

)

−

1

2

→∞. Fig. 1.9 shows the graph of ∆t/∆s vs. v.

Exercise 1.6. Investigate the

Fig. 1.9: ∆t/∆s = (1 − v

2

)

−

1

2

.

Doppler redshift. Let a source of light

pulses move with velocity v directly

away from an observer at rest in an iner-

tial frame. Let ∆t

e

be the time between

the emission of pulses, and ∆t

o

be the

time between the reception of pulses by

the observer.

a. Show that ∆t

o

= ∆t

e

+v∆t

e

/c .

b. Ignore time dilation in Eq. (1.6)

by setting ∆s = ∆t. Show that z = v/c

in this approximation.

c. Show that with c = 1, z = [(1 +

v)/(1 −v)]

1

2

−1 . Use the result of part

a.

Spacelike separated events. This case can be omitted without loss of

continuity, as it plays only a minor role in this book.

If E and F are not lightlike separated ([∆x[ = [∆t[) , or timelike separated

([∆x[ < [∆t[) , then [∆x[ > [∆t[ . Thus [∆t/∆x[ < 1 .

Fig. 1.10: ∆s

2

= ∆t

2

− ∆x

2

for spacelike separated events E and F. See the text.

For convenience, let E have coordinates E(0, 0). Fig. 1.10 shows the world-

line W

of an inertial observer O

**moving with velocity v = ∆t/∆x in I. (That
**

is not a typo.) L± are the light worldlines through F. Since c = 1 in I, the slope

of L± is ±1. Solve simultaneously the equations for L− and W

to obtain the

coordinates R(∆x, ∆t). Similarly, the equations for L+ and W

give S(∆x, ∆t).

According to Eq. (1.11), the proper time, as measured by O

, between the

timelike separated events S and E, and between E and R, is (∆x

2

−∆t

2

)

1

2

. Since

the times are equal, E and F are, by deﬁnition, simultaneous in an inertial frame

I

in which O

is at rest.

21

1.3 The Metric Postulate

Since c = 1 in I

, the distance between E and F in I

is (∆x

2

−∆t

2

)

1

2

. This

is the length of a rod at rest in I

**with its ends simultaneously at E and F. By
**

deﬁnition, this is the proper distance [∆s [ between the events. This proves Eq.

(1.11) for spacelike separated events.

A calculation similar to Eq. (1.12) gives

[∆s [ = (1 −v

2

)

1

2

[∆x[. (1.13)

The proper distance between spacelike separated events is less than an inertial

frame distance: [∆s [ < [∆x[. (This is not length contraction, which we discuss

in Appendix 4.)

In summary, the metric postulate is a mathematical expression of three phys-

ical assertions: a universal light speed in inertial frames, a formula for proper

time in inertial frames, and a formula for proper distance in inertial frames.

Local Forms. The metric postulate for a planar frame, Eq. (1.9), gives only

the distance along a straight line between two points. The diﬀerential version

of Eq. (1.9) gives the distance ds between neighboring points along any curve:

The metric postulate for a ﬂat surface, local form

Let P and Q be neighboring points. Let ds be the distance between

them. Let the points have coordinate diﬀerences (dx, dy) in a planar

frame. Then

ds

2

= dx

2

+dy

2

.

Thus, if a curve is parameterized (x(p), y(p)), a ≤ p ≤ b, then

ds

2

=

¸

dx

dp

2

+

dy

dp

2

¸

dp

2

,

and the length of the curve is

s =

b

p=a

ds =

b

p=a

¸

dx

dp

2

+

dy

dp

2

¸1

2

dp .

Of course, diﬀerent curves between two points can have diﬀerent lengths.

Exercise 1.7. Calculate the circumference of the circle x = r cos θ, y =

r sin θ, 0 ≤ θ ≤ 2π.

22

1.3 The Metric Postulate

The metric postulate for an inertial frame Eq. (1.10) is concerned only with

times measured by inertial clocks. The diﬀerential version of Eq. (1.10) gives

the time ds measured by any clock between neighboring events on its worldline:

The metric postulate for a ﬂat spacetime, local form

Let E and F be neighboring events. If E and F are lightlike sep-

arated, let ds = 0. If the events are timelike separated, let ds be

the time between them as measured by any (inertial or noninertial)

clock. Let the events have coordinate diﬀerences (dt, dx, dy, dz) in

an inertial frame. Then

ds

2

= dt

2

−dx

2

−dy

2

−dz

2

. (1.14)

From Eq. (1.14), if the worldline of a clock is parameterized

(t(p), x(p), y(p), z(p)), a ≤ p ≤ b,

then the time s to traverse the worldline, as measured by the clock, is

s =

b

p=a

ds =

b

p=a

¸

dt

dp

2

−

dx

dp

2

−

dy

dp

2

−

dz

dp

2

¸1

2

dp.

In general, clocks on diﬀerent worldlines between two events will measure dif-

ferent times between the events.

Exercise 1.8. Let a clock move between two events with a time diﬀerence

∆t. Let v be the small constant speed of the clock. Show that ∆t−∆s ≈

1

2

v

2

∆t.

Exercise 1.9. Consider a simpliﬁed Hafele-Keating experiment. One clock

remains on the ground and the other circles the equator in an airplane to the

west - opposite to the Earth’s rotation. Assume that the Earth is spinning on

its axis at one revolution per 24 hours in an inertial frame. (Thus the clock on

the ground is not at rest.)

Notation: ∆t is the duration of the trip in the inertial frame. v

g

is the

velocity of the clock on the ground and ∆s

g

is the time it measures for the trip.

v

a

and ∆s

a

are deﬁned similarly for the airplane.

Use Exercise 1.8 for each clock to show that the diﬀerence between the clocks

due to time dilation is

∆s

a

−∆s

g

=

1

2

(v

2

a

−v

2

g

)∆t.

Suppose that ∆t = 40 hours and the speed of the airplane with respect to

the ground is 1000 km/hr. Substitute values to obtain ∆s

a

−∆s

g

= 1.410

−7

s.

23

1.3 The Metric Postulate

Experimental Evidence. Because general relativistic eﬀects play a part in

the Hafele-Keating experiment (see Exercise 2.1), and because the uncertainty

of the experiment is large (±10%), this experiment is not a precision test of time

dilation for clocks. Much better evidence comes from observations of subatomic

particles called muons. When at rest the average lifetime of a muon is 3 10

−6

sec. According to the diﬀerential version of Eq. (1.12), if the muon is moving

in a circle with constant speed v, then its average life, as measured in the

laboratory, should be larger by a factor (1 − v

2

)

−

1

2

. An experiment performed

in 1977 showed this within an experimental error of .2%. In the experiment

v = .9994, giving a time dilation factor ∆t/∆s = (1 −v

2

)

−

1

2

= 29! The circular

motion was accompanied by an acceleration of 10

21

cm/sec

2

, and so this is a

test of the local form Eq. (1.14) of the metric postulate as well as the original

form Eq. (1.10).

There is excellent evidence for a universal light speed. First of all, realize

that if clocks at P and Q are synchronized according to the deﬁnition Eq.

(1.2), then the speed of light from P to Q is equal to the speed from Q to P.

We emphasize that with our deﬁnition of synchronized clocks this equality is a

matter of deﬁnition which can be neither conﬁrmed nor refuted by experiment.

The speed c of light can be measured by sending a pulse of light from a point

P to a mirror at a point Q at distance D and measuring the elapsed time 2T

for it to return. Then c = 2D/2T; c is a two way speed, measured with a single

clock. Equation 1.5 shows that this two way speed is equal to the one way speed

from P to Q. Thus the one way speed of light can be measured by measuring

the two way speed.

In a famous experiment, performed in 1887, A. A. Michelson and E. W.

Morley compared the two way speed of light in perpendicular directions from a

given point. Their experiment has been repeated many times, most accurately

by G. Joos in 1930, who found that any diﬀerence in the two way speeds is

less than six parts in 10

12

. The Michelson-Morley experiment is described in

Appendix 5. A modern version of the experiment using lasers was performed

in 1979 by A. Brillit and J. L. Hall. They found that any diﬀerence in the two

way speed of light in perpendicular directions is less than four parts in 10

15

.

See Appendix 6.

Another experiment, performed by R. J. Kennedy and E. M. Thorndike in

1932, found the two way speed of light to be the same, within six parts in 10

9

,

on days six months apart, during which time the Earth moved to the opposite

side of its orbit. See Appendix 7. Inertial frames in which the Earth is at rest

on days six months apart move with a relative speed of 60 km/sec (twice the

Earth’s orbital speed). A more recent experiment by D. Hils and Hall improved

the result by over two orders of magnitude.

These experiments provide good evidence that the two way speed of light

is the same in diﬀerent directions, places, and inertial frames and at diﬀerent

times. They thus provide strong motivation for our deﬁnition of synchronized

clocks: If the two way speed of light has always the same value, what could be

more natural than to deﬁne synchronized clocks by requiring that the one way

speed have this value?

24

1.3 The Metric Postulate

In all of the above experiments, the source of the light is at rest in the

inertial frame in which the light speed is measured. If light were like baseballs,

then the speed of a moving source would be imparted to the speed of light it

emits. Strong evidence that this is not so comes from observations of certain

neutron stars which are members of a binary star system and which emit X-ray

pulses at regular intervals. These systems are described in Sec. 3.1. If the

speed of the star were imparted to the speed of the X-rays, then various strange

eﬀects would be observed. For example, X-rays emitted when the neutron star

is moving toward the Earth could catch up with those emitted earlier when it

was moving away from the Earth, and the neutron star could be seen coming

and going at the same time! See Fig. 1.11.

Fig. 1.11: The speed of light is independent of the

speed of its source.

This does not happen; an

analysis of the arrival times

of the pulses at Earth made

in 1977 by K. Brecher shows

that no more than two parts

in 10

9

of the speed of the

source is added to the speed

of the X-rays. (It is not

possible to “see” the neutron

star in orbit around its com-

panion directly. The speed

of the neutron star toward or

away from the Earth can be determined from the Doppler redshift of the time

between pulses. See Exercise 1.6.)

Finally, recall from above that the universal light speed statement of the

metric postulate implies the statements about timelike and spacelike separated

events. Thus the evidence for a universal light speed is also evidence for the

other two statements.

The universal nature of the speed of light makes possible the modern deﬁni-

tion of the unit of length: “The meter is the length of the path traveled by light

during the time interval of 1/299,792,458 of a second.” Thus, by deﬁnition, the

speed of light is 299,792,458 m/sec.

25

1.4 The Geodesic Postulate

1.4 The Geodesic Postulate

We will ﬁnd it convenient to use superscripts to distinguish coordinates. Thus

we use (x

1

, x

2

) instead of (x, y) for planar frame coordinates.

Fig. 1.12: A geodesic in a planar frame.

Surface dwellers ﬁnd that straight lines in their surface can be parameterized

x

i

(s) = a

i

s + b

i

, i = 1, 2. The line in Fig. 1.12 can be parameterized by the

(proper) distance s from (b

1

, b

2

) to (x

1

, x

2

) : x

1

(s) = (cos θ)s + b

1

, x

2

(s) =

(sin θ)s +b

2

. Diﬀerentiate twice with respect to s to obtain

The geodesic postulate for a ﬂat surface

Parameterize a straight line with areclength s . Then in every planar

frame

¨ x

i

= 0, i = 1, 2. (1.15)

The straight lines are called geodesics.

Not all parameterizations of a straight line satisfy the geodesic diﬀerential

equations Eq. (1.15). Example: x

i

(p) = a

i

p

3

+b

i

.

26

1.4 The Geodesic Postulate

We will ﬁnd it convenient to use (x

0

, x

1

, x

2

, x

3

) instead of (t, x, y, z) for

inertial frame coordinates. Our third postulate for special relativity says that

inertial particles and light pulses move in a straight line at constant speed in an

inertial frame, i.e., their equations of motion are

x

i

= a

i

x

0

+b

i

, i = 1, 2, 3. (1.16)

(Diﬀerentiate to give dx

i

/dx

0

= a

i

; the velocity is constant.) For inertial parti-

cles this is called Newton’s ﬁrst law.

Set x

0

= p, a parameter; a

0

= 1; and b

0

= 0, and ﬁnd that worldlines of

inertial particles and light can be parameterized

x

i

(p) = a

i

p +b

i

, i = 0, 1, 2, 3 (1.17)

in an inertial frame. Eq. (1.17), unlike Eq. (1.16), is symmetric in all four

coordinates of the inertial frame. Also, Eq. (1.17) shows that the worldline is

a straight line in the spacetime. Thus “straight in spacetime” includes both

“straight in space” and “straight in time” (constant speed). See Exercise 1.1.

The worldlines are called geodesics.

Exercise 1.10. Eq. (1.17) parameterizes the worldline of an inertial parti-

cle with x

0

. Show the worldline can also parameterized with s, the proper time

along the worldline.

The geodesic postulate for a ﬂat spacetime

Worldlines of inertial particles and pulses of light can be parameter-

ized so that in every inertial frame

¨ x

i

= 0, i = 0, 1, 2, 3. (1.18)

For inertial particles we may take the parameter to be s.

The geodesic postulate is a mathematical expression of our physical assertion

that inertial particles and light move in a straight line at constant speed in an

inertial frame.

Exercise 1.11. Make as long a list as you can of analogous properties of

ﬂat surfaces and ﬂat spacetimes.

27

Chapter 2

Curved Spacetimes

2.1 History of Theories of Gravity

Recall the analogy from Chapter 1: A curved spacetime is to a ﬂat spacetime

as a curved surface is to a ﬂat surface. We explored the analogy between a ﬂat

surface and a ﬂat spacetime in Chapter 1. In this chapter we generalize from ﬂat

surfaces and ﬂat spacetimes (spacetimes without signiﬁcant gravity) to curved

surfaces and curved spacetimes (spacetimes with signiﬁcant gravity). General

relativity interprets gravity as a curvature of spacetime.

Fig. 2.1: The position of a planet

( ◦ ) with respect to the stars

changes nightly.

Before embarking on a study of gravity

in general relativity let us brieﬂy review the

history of theories of gravity. These theo-

ries played a major role in the rise of sci-

ence. Theories of gravity have their roots in

attempts of the ancients to predict the mo-

tion of the planets. The position of a planet

with respect to the stars changes from night

to night. Some exhibit a “loop” motion, as

in Fig. 2.1. The word “planet” is from the

Greek “planetai”: wanderers.

In the second century, Claudius Ptolemy

devised a scheme to explain these motions. Ptolemy placed the Earth at the

center of the universe with a planet moving on a small circle, called an epicycle,

while the center of the epicycle moves uniformly on a larger circle, the defer-

ent, around the Earth. See Fig. 2.2. By appropriately choosing the radii of

the epicycle and deferent, as well as the speeds involved, Ptolemy was able to

reproduce, with fair accuracy, the motions of the planets. However, this did not

work exactly and so Ptolemy added epicycles to the epicycles. To explain the

motions of the Moon and the known planets, Ptolemy needed 77 epicycles! This

remarkable but cumbersome theory was accepted for over 1000 years.

28

2.1 History of Theories of Gravity

Fig. 2.2: Ptolemy’s theory of plane-

tary motion.

In 1543 Nicholas Copernicus published

a theory which was to revolutionize science

and our perception of our place in the uni-

verse: he placed the Sun near the center of

our planetary system. Copernicus retained

the system of uniform circular motion using

epicycles and deferents, but by placing the

Sun near the center he was able to reduce

greatly their number.

In 1609 Johannes Kepler published a

theory which discarded the whole system of

epicycles and deferents and replaced it with a simpler description: the path of a

planet is an ellipse with the Sun at one focus. At about the same time Galileo

Galilei was investigating the acceleration of objects near the Earth’s surface.

He found two things of interest for us: the acceleration is constant in time and

independent of the mass and composition of the falling object.

In 1687 Isaac Newton published a theory of grav-

Fig. 2.3: Falling and

orbital motion are the

same.

ity which explained Kepler’s astronomical and Galileo’s

terrestrial ﬁndings as manifestations of the same phe-

nomenon: gravity. To understand how orbital motion

is related to falling motion, refer to Fig. 2.3. The

curves A, B, C are the paths of objects leaving the top

of a tower with greater and greater horizontal veloci-

ties. They hit the ground farther and farther from the

bottom of the tower until C when the object goes into

orbit!

Mathematically, Newton’s theory says that a planet

in the Sun’s gravity or an apple in the Earth’s gravity

is pulled instantaneously by the central body (do not

ask how!), causing an acceleration

a = −

κM

r

2

, (2.1)

where κ is the Newtonian gravitational constant, M is the mass of the central

body, and r is the distance to the center of the central body. Eq. (2.1) implies

that the planets orbit the Sun in ellipses, in accord with Kepler’s ﬁndings. See

Appendix 8. By taking the distance r to the Earth’s center to be sensibly

constant near the Earth’s surface, we see that Eq. (2.1) is also in accord with

Galileo’s ﬁndings: a is constant in time and is independent of the mass and

composition of the falling object.

Newton’s theory has enjoyed enormous success. A spectacular example oc-

curred in the nineteenth century. Observations of the position of the planet

Uranus disagreed with the predictions of Newton’s theory of gravity, even after

taking into account the gravitational eﬀects of the other known planets. The

discrepancy was about 4 arcminutes – 1/8

th

of the angular diameter of the

29

2.1 History of Theories of Gravity

moon. In 1846, U. LeVerrier, a French mathematician, calculated that a new

planet, beyond Uranus, could account for the discrepancy. He wrote J. Galle, an

astronomer at the Berlin observatory, telling him where the new planet should

be – and Neptune was discovered! It was within 1 arcdegree of LeVerrier’s

prediction.

Even today, calculations of spacecraft trajectories are made using Newton’s

theory. The incredible accuracy of his theory will be examined further in Sec.

3.3.

Nevertheless, Einstein rejected Newton’s theory because it is based on pre-

relativity ideas about time and space which, as we have seen, are not correct.

For example, the acceleration in Eq. (2.1) is instantaneous with respect to a

universal time.

30

2.2 The Key to General Relativity

2.2 The Key to General Relativity

A curved surface is diﬀerent from a ﬂat surface. However, a simple observation

by the nineteenth century mathematician Karl Friedrich Gauss provides the key

to the construction of the theory of surfaces: a small region of a curved surface

is very much like a small region of a ﬂat surface. This is familiar: a small

region of a (perfectly) spherical Earth appears ﬂat. This is why it took so long

to discover that it is not ﬂat. On an apple, a smaller region must be chosen

before it appears ﬂat.

In the next three sections we shall formalize Gauss’ observation by taking

the three postulates for ﬂat surfaces from Chapter 1, restricting them to small

regions, and then using them as postulates for curved surfaces.

We shall see that a curved spacetime is diﬀerent from a ﬂat spacetime.

However, a simple observation of Einstein provides the key to the construction

of general relativity: a small region of a curved spacetime is very much like a

small region of a ﬂat spacetime. To understand this, we must extend the concept

of an inertial object to curved spacetimes.

Passengers in an airplane at rest on the ground or ﬂying in a straight line

at constant speed feel gravity, which feels very much like an inertial force. Ac-

celerometers respond to the gravity. On the other hand, astronauts in orbit or

falling radially toward the Earth feel no inertial forces, even though they are

not moving in a straight line at constant speed with respect to the Earth. And

an accelerometer carried by the astronauts will register zero. We shall include

gravity as an inertial force and, as in special relativity, call an object inertial

if it experiences no inertial forces. An inertial object in gravity is in free fall.

Forces other than gravity do not act on it.

We now rephrase Einstein’s observation: as viewed by inertial observers, a

small region of a curved spacetime is very much like a small region of a ﬂat

spacetime. We see this vividly in motion pictures of astronauts in orbit. No

gravity is apparent in their cabin; objects suspended at rest remain at rest.

Inertial objects in the cabin move in a straight line at constant speed, just as

in a ﬂat spacetime. The Newtonian theory predicts this: according to Eq. 2.1

an inertial object and the cabin accelerate the same with respect to the Earth

and so they do not accelerate with respect to each other.

In the next three sections we shall formalize Einstein’s observation by tak-

ing our three postulates for ﬂat spacetimes, restricting them to small spacetime

regions, and then using them as our ﬁrst three (of four) postulates for curved

spacetimes. The local inertial frame postulate asserts the existence of small in-

ertial cubical lattices with synchronized clocks to serve as coordinate systems

in small regions of a curved spacetime. The metric postulate asserts a universal

light speed and a slowing of moving clocks in local inertial frames. The geodesic

postulate asserts that inertial particles and light move in a straight line at con-

stant speed in local inertial frames. We ﬁrst discuss experimental evidence for

the three postulates.

31

2.2 The Key to General Relativity

Experiments of R. Dicke and of V. B. Braginsky, performed in the 1960’s,

verify to extraordinary accuracy Galileo’s ﬁnding incorporated into Newton’s

theory: the acceleration of a free falling object in gravity is independent of its

mass and composition. (See Sec. 2.1.) We may reformulate this in the language

of spacetimes: the worldline of an inertial object in a curved spacetime is inde-

pendent of its mass and composition. The geodesic postulate will incorporate

this by not referring to the mass or composition of the inertial objects whose

worldlines it describes.

Dicke and Braginsky used the Sun’s

Fig. 2.4: Masses A and B accelerate

the same toward the Sun.

gravity. We can understand the principle

of their experiments from the simpliﬁed di-

agram in Fig. 2.4. The weights A and B,

supported by a quartz ﬁber, are, with the

Earth, in free fall around the Sun. Dicke

and Braginsky used various substances with

various properties for the weights. Any dif-

ference in their acceleration toward the Sun

would cause a twisting of the ﬁber. Due to

the Earth’s rotation, the twisting would be

in the opposite direction twelve hours later.

The apparatus had a resonant period of oscillation of 24 hours so that oscillations

could build up. In Braginsky’s experiment the diﬀerence in the acceleration of

the weights toward the Sun was no more than one part in 10

12

of their mu-

tual acceleration toward the Sun. A planned satellite test (STEP) will test the

equality of accelerations to one part in 10

18

.

A related experiment shows that the Earth and the Moon accelerate the same

in the Sun’s gravity, despite the huge diﬀerence in their masses. If this were

not so, then there would be unexpected changes in the Earth-Moon distance.

Changes in this distance can be measured within 2 cm (!) by timing the return

of a laser pulse sent from Earth to mirrors on the Moon left by astronauts.

This is part of the lunar laser experiment. The measurements show that the

relative acceleration of the Earth and Moon is no more than a part in 10

13

of

their mutual acceleration toward the Sun. The experiment also shows that the

Newtonian gravitational constant κ in Eq. (2.1) does not change by more than

1 part in 10

12

per year. The constant also appears in Einstein’s ﬁeld equation,

Eq. (2.27).

There is another diﬀerence between the Earth and the Moon which might

cause a diﬀerence in their acceleration toward the Sun. Imagine disassembling

to Earth into small pieces and separating the pieces far apart. The separation

requires energy input to counteract the gravitational attraction of the pieces.

This energy is called the Earth’s gravitational binding energy . By Einstein’s

principle of the equivalence of mass and energy (E = mc

2

), this energy is

equivalent to mass. Thus the separated pieces have more total mass than the

Earth. The diﬀerence is small – 4.6 parts in 10

10

. It is 23 times smaller for the

Moon. Since the gravitational binding energy mass is in some sense diﬀerent

from “ordinary” mass, one can wonder whether this diﬀerence between the Earth

32

2.2 The Key to General Relativity

and Moon causes a diﬀerence in their acceleration toward the Sun. The lunar

laser experiment shows that this does not happen. This is something that the

Dicke and Braginsky experiments cannot test.

The last experiment we shall consider as evidence for the three postulates is

the terrestrial redshift experiment. It was ﬁrst performed by R. V. Pound and G.

A. Rebka in 1960 and then more accurately by Pound and J. L. Snider in 1964.

The experimenters put a source of gamma radiation at the bottom of a tower.

Radiation received at the top of the tower was redshifted: z = 2.5 10

−15

,

within an experimental error of about 1%. This is a gravitational redshift.

According to the discussion following Eq. (1.6), an observer at the top of

the tower would see a clock at the bottom run slowly. Clocks at rest at diﬀerent

heights in the Earth’s gravity run at diﬀerent rates! Part of the result of the

Hafele-Keating experiment is due to this. See Exercise 2.1.

We showed in Sec. 1.3 that the assumption Eq. (1.4), necessary for synchro-

nizing clocks at rest in the coordinate lattice of an inertial frame, is equivalent

to a zero redshift between the clocks. This assumption fails for clocks at the

top and bottom of the tower. Thus clocks at rest in a small coordinate lattice

on the ground cannot be (exactly) synchronized.

We now show that the experiment provides evidence that clocks at rest in

a small inertial lattice can be synchronized. In the experiment, the tower has

(upward) acceleration g, the acceleration of Earth’s gravity, in a small inertial

lattice falling radially toward Earth. We will show shortly that the same redshift

would be observed with a tower having acceleration g in an inertial frame in

a ﬂat spacetime. This is another example of small regions of ﬂat and curved

spacetimes being alike. Thus it is reasonable to assume that there would be

no redshift with a tower at rest in a small inertial lattice in gravity, just as

with a tower at rest in an inertial frame. (It is desirable to test this directly

by performing the experiment in orbit.) In this way, the experiment provides

evidence that the condition Eq. (1.4), necessary for clock synchronization, is

valid for clocks at rest in a small inertial lattice. Loosely speaking, we may say

that since light behaves “properly” in a small inertial lattice, light accelerates

the same as matter in gravity.

We now calculate the Doppler redshift for a tower with acceleration g in an

inertial frame. Suppose the tower is momentarily at rest when gamma radiation

is emitted. The radiation travels a distance h, the height of the tower, in the

inertial frame. (We ignore the small distance the tower moves during the ﬂight

of the radiation. We shall also ignore the time dilation of clocks in the moving

tower and the length contraction – see Appendix 4 – of the tower. These eﬀects

are far too small to be detected by the experiment.) Thus the radiation takes

time t = h/c to reach the top of the tower. (For clarity we do not take c = 1.)

In this time the tower acquires a speed v = gt = gh/c in the inertial frame.

From Exercise 1.6, this speed causes a Doppler redshift

z =

v

c

=

gh

c

2

. (2.2)

In the experiment, h = 2250 cm. Substituting numerical values in Eq.

33

2.2 The Key to General Relativity

2.2 gives the value of z measured in the terrestrial redshift experiment; the

gravitational redshift for towers accelerating in inertial frames is the same as

the Doppler redshift for towers accelerating in small inertial lattices near Earth.

Exercise 3.4 shows that a rigorous calculation in general relativity also gives Eq.

(2.2).

Exercise 2.1. Let h be the height at which the airplane ﬂies in the simpliﬁed

Hafele-Keating experiment of Exercise 1.9. Show that the diﬀerence between

the clocks due to the gravitational redshift is

∆s

a

−∆s

g

= gh∆t.

Suppose that h = 10 km. Substitute values to obtain ∆s

a

−∆s

g

= 1.610

−7

sec.

Adding this to the time dilation diﬀerence of Exercise 1.9 gives a diﬀerence

of 3.0 10

−7

sec. Exercise 3.3 shows that a rigorous calculation in general

relativity gives the same result.

The Global Positioning System (GPS) is a system of 24 Earth satellites.

Using this system, a user on Earth can determine their position to within a few

meters with a hand held device. The system must take into account the special

and general relativistic eﬀects of Exercise 2.1 to function properly. In fact, the

eﬀects are 10,000 times too large to be ignored.

34

2.3 The Local Inertial Frame Postulate

2.3 The Local Inertial Frame Postulate

Fig. 2.5: A local planar frame.

Suppose that curved surface dwellers attempt to

construct a square coordinate grid using rigid

rods constrained (of course) to their surface. If

the rods are short enough, then at ﬁrst they will

ﬁt together well. But, owing to the curvature of

the surface, as the grid gets larger the rods must

be forced a bit to connect them. This will cause

stresses in the lattice and it will not be quite

square. Surface dwellers call a small (nearly)

square coordinate grid a local planar frame at P,

where P is the point at the origin of the grid. In

smaller regions around P, the grid must become

more square. See Fig. 2.5.

The local planar frame postulate for a curved surface

A local planar frame can be constructed at any point P of a curved

surface with any orientation.

We shall see that local planar frames

Fig. 2.6: Spherical coordinates (φ, θ)

on a sphere.

at P provide surface dwellers with an in-

tuitive description of properties of a curved

surface at P. However, in order to study

the surface as a whole, they need global co-

ordinates, deﬁned over the entire surface.

There are, in general, no natural global co-

ordinate systems to single out in a curved

surface as planar frames are singled out in a

ﬂat surface. Thus they attach global coordi-

nates (y

1

, y

2

) in an arbitrary manner. The

only restrictions are that diﬀerent points

must have diﬀerent coordinates and nearby

points must receive nearby coordinates. In general, the coordinates will have

no geometric meaning; they merely serve to label the points of the surface.

One common way for us (but not surface dwellers) to attach global coordi-

nates to a curved surface is to parameterize it in three dimensional space:

x = x(y

1

, y

2

), y = y(y

1

, y

2

), z = z(y

1

, y

2

) . (2.3)

As (y

1

, y

2

) varies, (x, y, z) varies on the surface. Assign coordinates (y

1

, y

2

) to

the point (x, y, z) on the surface given by Eq. (2.3). For example, Fig. 2.6 shows

spherical coordinates (y

1

, y

2

) = (φ, θ) on a sphere of radius R. The coordinates

are assigned by the usual parameterization

x = Rsin φcos θ, y = Rsin φsin θ, z = Rcos φ. (2.4)

35

2.3 The Local Inertial Frame Postulate

Surface dwellers on the sphere could assign these coordinates without leaving

the surface. For example, the circle of latitude at distance r from the North

Pole is assigned the coordinate φ = r/R, where 2πR is the distance around a

great circle. See Fig. 2.10. (We shall see in Sec. 2.6 that surface dwellers can

determine great circles and R without leaving the surface.)

The global coordinate postulate for a curved surface

The points of a curved surface can be labeled with coordinates

(y

1

, y

2

).

(Technically, the postulate should state that a curved surface is a two dimen-

sional manifold. The statement given will suﬃce for us.)

In the last section we saw that inertial objects in an astronaut’s cabin behave

as if no gravity were present. Actually, they will not behave ex:actly as if no

gravity were present. To see this, assume for simplicity that their cabin is falling

radially toward Earth. Inertial objects in the cabin do not accelerate exactly the

same with respect to the Earth because they are at slightly diﬀerent distances

and directions from the Earth’s center. See Fig. 2.7. Thus, an object initially

at rest near the top of the cabin will slowly separate from one initially at rest

near the bottom. In addition, two objects initially at rest at the same height

will slowly move toward each other as they both fall toward the center of the

Earth. These changes in velocity are called tidal accelerations. (Why?) They

are caused by small diﬀerences in the Earth’s gravity at diﬀerent places in the

cabin. They become smaller in smaller regions of space and time, i.e., in smaller

regions of spacetime.

Suppose that astronauts in a curved spacetime at-

Fig. 2.7: Tidal accel-

erations in a radially

free falling cabin.

tempt to construct an inertial cubical lattice using rigid

rods. If the rods are short enough, then at ﬁrst they will

ﬁt together well. But as the grid gets larger, the lattice

will have to resist tidal accelerations, and the rods can-

not all be inertial. This will cause stresses in the lattice

and it will not be quite cubical.

In the last section we saw that the terrestrial redshift

experiment provides evidence that clocks in a small iner-

tial lattice can be synchronized. Actually, due to small

diﬀerences in the gravity at diﬀerent places in the lattice,

an attempt to synchronize the clocks with the one at the

origin with the procedure of Sec. 1.3 will not quite work. However, we can hope

that the procedure will work with as small an error as desired by restricting the

lattice to a small enough region of a spacetime.

A small (nearly) cubical lattice with (nearly) synchronized clocks is called a

local inertial frame at E, where E is the event at the origin of the lattice when

the clock there reads zero. In smaller regions around E, the lattice is more

cubical and the clocks are more nearly synchronized. Local inertial frames are

in free fall.

36

2.3 The Local Inertial Frame Postulate

The local inertial frame postulate for a curved spacetime

A local inertial frame can be constructed at any event E of a curved

spacetime, with any orientation, and with any inertial object at E

instantaneously at rest in it.

We shall ﬁnd that local inertial frames at E provide an intuitive description

of properties of a curved spacetime at E. However, in order to study a curved

spacetime as a whole, we need global coordinates, deﬁned over the entire space-

time. There are, in general, no natural global coordinates to single out in a

curved spacetime, as inertial frames were singled out in a ﬂat spacetime. Thus

we attach global coordinates in an arbitrary manner. The only restrictions are

that diﬀerent events must receive diﬀerent coordinates and nearby events must

receive nearby coordinates. In general, the coordinates will have no physical

meaning; they merely serve to label the events of the spacetime.

Often one of the coordinates is a “time” coordinate and the other three are

“space” coordinates, but this is not necessary. For example, in a ﬂat spacetime

it is sometimes useful to replace the coordinates (t, x) with (t +x, t −x).

The global coordinate postulate for a curved spacetime

The events of a curved spacetime can be labeled with coordinates

(y

0

, y

1

, y

2

, y

3

).

In the next two sections we give the metric and geodesic postulates of gen-

eral relativity. We ﬁrst express the postulates in local inertial frames. This

local form of the postulates gives them the same physical meaning as in special

relativity. We then translate the postulates to global coordinates. This global

form of the postulates is unintuitive and complicated but is necessary to carry

out calculations in the theory.

We can use arbitrary global coordinates in ﬂat as well as curved spacetimes.

We can then put the metric and geodesic postulates of special relativity in the

same global form that we shall obtain for these postulates for curved space-

times. We do not usually use arbitrary coordinates in ﬂat spacetimes because

inertial frames are so much easier to use. We do not have this luxury in curved

spacetimes.

It is remarkable that we shall be able to describe curved spacetimes intrin-

sically, i.e., without describing it as curved in a higher dimensional ﬂat space.

Gauss created the mathematics necessary to describe curved surfaces intrinsi-

cally in 1827. G. B. Riemann generalized Gauss’ mathematics to curved spaces

of higher dimension in 1854. His work was extended by several mathematicians.

Thus the mathematics necessary to describe curved spacetimes intrinsically was

waiting for Einstein when he needed it.

37

2.4 The Metric Postulate

2.4 The Metric Postulate

Fig. 2.5 shows that local planar frames provide curved surface dwellers with a

convenient way to express inﬁnitesimal distances on a curved surface.

The metric postulate for a curved surface, local form

Let point Q have coordinates (dx

1

, dx

2

) in a local planar frame at

P. Let ds be the distance between the points. Then

ds

2

= (dx

1

)

2

+ (dx

2

)

2

. (2.5)

Even though the local planar frame extends a ﬁnite distance from P, Eq. (2.5)

holds only for inﬁnitesimal distances from P.

We now express Eq. (2.5) in terms of global coordinates. Set the matrix

f

◦

= (f

◦

mn

) =

1 0

0 1

.

Then Eq. (2.5) can be written

ds

2

=

2

¸

m,n=1

f

◦

mn

dx

m

dx

n

. (2.6)

Henceforth we use the Einstein summation convention by which an index which

appears twice in a term is summed without using a Σ. Thus Eq. (2.6) becomes

ds

2

= f

◦

mn

dx

m

dx

n

. (2.7)

As another example of the summation convention, consider a function f(y

1

, y

2

).

Then we may write the diﬀerential df = (∂f/∂y

i

) dy

i

.

Let P and Q be neighboring points on a curved surface with coordinates

(y

1

, y

2

) and (y

1

+ dy

1

, y

2

+ dy

2

) in a global coordinate system. Let Q have

coordinates (dx

1

, dx

2

) in a local planar frame at P. We may think of the (x

i

)

coordinates as functions of the (y

j

) coordinates, just as cartesian coordinates in

the plane are functions of polar coordinates: x = r cos θ, y = r sin θ. This gives

meaning to the partial derivatives ∂x

i

/∂y

j

. From Eq. (2.7), the distance from

P to Q is

ds

2

= f

◦

mn

dx

m

dx

n

= f

◦

mn

∂x

m

∂y

j

dy

j

∂x

n

∂y

k

dy

k

(sumon m, n, j, k)

=

f

◦

mn

∂x

m

∂y

j

∂x

n

∂y

k

dy

j

dy

k

= g

jk

(y) dy

j

dy

k

, (2.8)

38

2.4 The Metric Postulate

where we have set

g

jk

(y) = f

◦

mn

∂x

m

∂y

j

∂x

n

∂y

k

. (2.9)

Since (f

◦

mn

) is symmetric, so is (g

jk

). Use a local planar frame at each point of

the surface in this manner to translate the local form of the metric postulate,

Eq. (2.5), to global coordinates:

The metric postulate for a curved surface, global form

Let (y

i

) be global coordinates on the surface. Let ds be the distance

between neighboring points (y

i

) and (y

i

+ dy

i

). Then there is a

symmetric matrix (g

jk

(y

i

)) such that

ds

2

= g

jk

(y

i

) dy

j

dy

k

. (2.10)

(g

jk

(y

i

)) is called metric of the surface with respect to the coordinates (y

i

).

Exercise 2.2. Show that the metric for the (φ, θ) coordinates on the sphere

in Eq. (2.4) is

R

2

0

0 R

2

sin

2

φ

. (2.11)

Do this in two ways:

a. By converting from the metric of a local planar frame. Show that for

a local planar frame whose x

1

-axis coincides with a circle of latitude, dx

1

=

Rsinφdθ and dx

2

= Rdφ. See Fig. 2.10.

b. Use Eq. (2.4) to convert ds

2

= dx

2

+dy

2

+dz

2

to (φ, θ) coordinates.

Exercise 2.3. Consider the hemisphere z = (R

2

− x

2

− y

2

)

1

2

. Assign

coordinates (x, y) to the point (x, y, z) on the hemisphere. Find the metric in this

coordinate system. Express your answer as a matrix. Hint: Use z

2

= R

2

−x

2

−y

2

to compute dz

2

.

We should not think of a vector as its components (v

i

) , but as a single object

v which represents a magnitude and direction (an arrow). In a given coordinate

system the vector acquires components. The components will be diﬀerent in

diﬀerent coordinate systems.

Similarly, we should not think of the metric as its components (g

jk

) , but as

a single object g which represents inﬁnitesimal distances. In a given coordinate

system the metric acquires components. The components will be diﬀerent in

diﬀerent coordinate systems.

Exercise 2.4. Let (y

i

) and (¯ y

i

) be two coordinate systems on the same

surface, with metrics (g

jk

(y

i

)) and (¯ g

pq

(¯ y

i

)). Show that

¯ g

pq

= g

jk

∂y

j

∂¯ y

p

∂y

k

∂¯ y

q

. (2.12)

Hint: See Eq. (2.8).

39

2.4 The Metric Postulate

The metric postulate for a curved spacetime, local form

Let event F have coordinates (dx

i

) in a local inertial frame at E.

If E and F are lightlike separated, let ds = 0. If the events are

timelike separated, let ds be the time between them as measured by

any (inertial or noninertial) clock. Then

ds

2

= (dx

0

)

2

−(dx

1

)

2

−(dx

2

)

2

−(dx

3

)

2

. (2.13)

The metric postulate asserts a universal light speed and a formula for proper

time in local inertial frames. (See the discussion following Eq. (1.10).) This

is another instance of the key to general relativity: a small region of a curved

spacetime is very much like a small region of a ﬂat spacetime.

We now translate the metric postulate to global coordinates. Eq. (2.13) can

be written

ds

2

= f

◦

mn

dx

m

dx

n

, (2.14)

where

f

◦

= (f

◦

mn

) =

¸

¸

¸

1 0 0 0

0 −1 0 0

0 0 −1 0

0 0 0 −1

¸

.

Using Eq. (2.14), the calculation Eq. (2.8), which produced the global form

of the metric postulate for curved surfaces, now produces the global form of the

metric postulate for curved spacetimes.

The metric postulate for a curved spacetime, global form

Let (y

i

) be global coordinates on the spacetime. If neighboring

events (y

i

) and (y

i

+ dy

i

) are lightlike separated, let ds = 0. If

the events are timelike separated, let ds be the time between them

as measured by a clock moving between them. Then there is a sym-

metric matrix (g

jk

(y

i

)) such that

ds

2

= g

jk

(y

i

) dy

j

dy

k

. (2.15)

(g

jk

(y

i

)) is called the metric of the spacetime with respect to (y

i

).

40

2.5 The Geodesic Postulate

2.5 The Geodesic Postulate

Curved surface dwellers ﬁnd that some curves in their surface are straight in

local planar frames. They call these curves geodesics. Fig. 2.8 shows that the

equator is a geodesic but the other circles of latitude are not. To traverse a

geodesic, a surface dweller need only always walk “straight ahead”. A geodesic

is as straight as possible, given that it is constrained to the surface.

As with the geodesic postulate for ﬂat surfaces Eq. (1.15), we have

The geodesic postulate for a curved surface, local form

Parameterize a geodesic with areclength s . Let point P be on the

worldline. Then in every local planar frame at P

¨ x

i

(P) = 0, i = 1, 2. (2.16)

We now translate Eq. (2.16) to global coordinates y to obtain the global

form of the geodesic equations. We ﬁrst need to know that the metric g = (g

ij

)

has an inverse g

−1

= (g

jk

).

Exercise 2.5. a. Let the matrix a = (∂x

n

/∂y

k

).

Fig. 2.8: The equator is

the only circle of latitude

which is a geodesic.

Show that the inverse matrix a

−1

= (∂y

k

/∂x

j

).

b. Show that Eq. (2.9) can be written g = a

t

f

◦

a,

where t means “transpose”.

c. Prove that g

−1

= a

−1

(f

◦

)

−1

(a

−1

)

t

.

Introduce the notation ∂

k

g

im

= ∂g

im

/∂y

k

. Deﬁne

the Christoﬀel symbols:

Γ

i

jk

=

1

2

g

im

[∂

k

g

jm

+∂

j

g

mk

−∂

m

g

jk

] . (2.17)

Note that Γ

i

jk

= Γ

i

kj

. The Γ

i

jk

, like the g

jk

, are

functions of the coordinates.

Exercise 2.6. Show that for the metric Eq. (2.11)

Γ

φ

θθ

= −sinφ cos φ and Γ

θ

θφ

= Γ

θ

φθ

= cot φ.

The remaining Christoﬀel symbols are zero.

You should not try to assign a geometric meaning to the Christoﬀel symbols;

they should simply be thought of as what appears when the geodesic postulate

is translated from its local form Eq. (2.16) (which has an evident geometric

interpretation) to its global form (which does not):

The geodesic postulate for a curved surface, global form

Parameterize a geodesic with areclength s . Then in every global

coordinate system

¨ y

i

+ Γ

i

jk

˙ y

j

˙ y

k

= 0, i = 1, 2. (2.18)

41

2.5 The Geodesic Postulate

Appendix 9 translates the local form of the postulate to the global form.

The translation requires an assumption. A local planar frame at P extends to a

ﬁnite region around P. Let f = (f

mn

(x)) represent the metric in this coordinate

system. According to Eq. (2.7), (f

mn

(P)) = f

◦

. Appendix 10 shows that there

are coordinates, called geodesic coordinates, satisfying this relationship and also

∂

i

f

mn

(P) = 0 (2.19)

for all m, n, i. A function with a zero derivative at a point is not changing much

at the point. In this sense Eq. (2.19) states that f stays close to f

◦

near P.

Since a local planar frame at P is constructed to approximate a planar frame

as closely as possible near P, surface dwellers assume that the metric of a local

planar frame at P satisﬁes Eq. (2.19).

Exercise 2.7. Show that the metric of Exercise 2.3 satisﬁes Eq. (2.19) at

(x, y) = (0, 0) .

Exercise 2.8. Show that Eq. (2.18) reduces to Eq. (2.16) for local planar

frames.

Exercise 2.9. Show that the equator is the only circle of latitude which is

a geodesic. Of course, all great circles on a sphere are geodesics. Use the result

of Exercise 2.6. Before using the geodesic equations you must parameterize the

circles with s.

The geodesic postulate for a curved spacetime, local form

Worldlines of inertial particles and pulses of light can be parame-

terized so that if E is on the worldline, then in every local inertial

frame at E

¨ x

i

(E) = 0, i = 0, 1, 2, 3. (2.20)

For inertial particles we may take the parameter to be s.

The worldlines are called geodesics.

The geodesic postulate asserts

Fig. 2.9: The path of an inertial particle in a

lattice stuck to the Earth and in an inertial

lattice. The dots are at equal time intervals.

that inertial particles and light

move in a straight line at constant

speed in a local inertial frame. (See

the remarks following Eq. (1.18).)

The worldline of an inertial particle

or pulse of light in a curved space-

time is as straight as possible (in

space and time – see the remarks

following Eq. (1.17)), given that it is constrained to the spacetime. The geodesic

is straight in local inertial frames, but looks curved when viewed in an “inap-

propriate” coordinate system. See Fig. 2.9.

Einstein’s “straightest worldline in a curved spacetime” description is very

diﬀerent from Newton’s “curved path in a ﬂat space” description.

42

2.5 The Geodesic Postulate

We now assume that the metric of a local inertial frame at E satisﬁes Eq.

(2.19) for the same reasons as given above for local planar frames. Then Ap-

pendix 9 translates the local form of the geodesic postulate for curved spacetimes

to the global form:

The geodesic postulate for a curved spacetime, global form

Worldlines of inertial particles and pulses of light can be parameter-

ized so that in every global coordinate system

¨ y

i

+ Γ

i

jk

˙ y

j

˙ y

k

= 0, i = 0, 1, 2, 3. (2.21)

For inertial particles we may take the parameter to be s.

43

2.6 The Field Equation

2.6 The Field Equation

Previous sections of this chapter explored the similarities between small regions

of ﬂat and curved surfaces and between small regions of ﬂat and curved space-

times. This section explores the diﬀerences.

The local forms of our curved surface postulates show

Fig. 2.10: K = 1/R

2

on a sphere.

that in many ways a small region of a curved surface is

like a small region of a ﬂat surface. Surface dwellers

might suppose that all diﬀerences between the regions

vanish as the regions become smaller. This is not so.

To see this, pass geodesics through a point P in every

direction.

Exercise 2.10. Show that there is a unique geodesic

through every point of a curved surface in every direction.

Hint: Use a basic theorem on the existence and unique-

ness of solutions of systems of diﬀerential equations.

Connect all the points at distance r from P along the

geodesics. For surface dwellers, r is the radius of a circle, C. Let C(r) be the

circumference of the circle. Deﬁne the curvature K of the surface at P:

K =

3

π

lim

r→0

2π r −C(r)

r

3

. (2.22)

Clearly, K = 0 for a ﬂat surface. From Fig. 2.10, C(r) < 2πr for a sphere

and so K ≥ 0. From Fig. 2.10, we ﬁnd

C(r) = 2πRsin φ = 2πRsin(r/R) = 2πR

r/R −(r/R)

3

/6 +. . .

.

A quick calculation shows that K = 1/R

2

. The cur-

Fig. 2.11: K = −1/R

2

on

a pseudosphere.

vature is a diﬀerence between regions of a sphere and

a ﬂat surface which does not vanish as the regions

become smaller.

The surface of revolution of Fig. 2.11 is a pseudo-

sphere. The horizontal “circles of latitude” are con-

cave inward and the vertical “lines of longitude” are

concave outward. Thus C(r) > 2πr and so K ≤ 0.

Exercise 2.14 shows that K = −1/R

2

, where R is a

constant. Despite the examples, in general K varies

from point to point in a curved surface.

Distances and geodesics are involved in the deﬁ-

nition of K. But distances determine geodesics: dis-

tances determine the metric, which determines the

Christoﬀel symbols Eq. (2.17), which determine the geodesics Eq. (2.18). Thus

we learn an important fact: distances determine K. Thus K is measurable by

surface dwellers.

Exercise 2.11. Show that a map of a region of the Earth must distort

distances. Take the Earth to be perfectly spherical. Make no calculations.

44

2.6 The Field Equation

We can roll a ﬂat piece of paper into a cylinder without distorting distances

along curves in the paper and thus without changing K. Thus K = 0 for

the cylinder. Viewed from the outside, the cylinder is curved, and so K = 0

seems “wrong”. However, viewed from within (and remember, we are describing

curved surfaces and spacetimes without reference to a higher dimensional space),

the rolling does not distort distances in the paper. Thus surface dwellers could

not detect the curvature seen from the outside. Thus K “should” be zero for

the cylinder.

The formula expressing K in terms of distances was given by Gauss. If

g

12

= 0, then

K = −(g

11

g

22

)

−

1

2

∂

1

g

−

1

2

11

∂

1

g

1

2

22

+∂

2

g

−

1

2

22

∂

2

g

1

2

11

¸

∂

i

≡ ∂/∂y

i

. (2.23)

Exercise 2.12. Show that Eq. (2.23) gives K = 1/R

2

for a sphere of radius

R. Use Eq. (2.11).

Exercise 2.13. Generate a surface of revolution by rotating the param-

eterized curve y = f(u), z = h(u) about the z-axis. Let (r, θ, z) be cylin-

drical coordinates and parameterize the surface with coordinates (u, θ). Use

ds

2

= dr

2

+r

2

dθ

2

+dz

2

to show that the metric is

f

2

+h

2

0

0 f

2

.

Exercise 2.14. If y = Re

−u

and z = R

u

0

(1 − e

−2t

)

1

2

dt, then the surface

of revolution in Exercise 2.13 is the pseudosphere of Fig. 2.11. Show that

K = −1/R

2

for the pseudosphere.

Exercise 2.15. If y = 1 and z = u, then the surface of revolution in

Exercise 2.13 is a cylinder. Show that K = 0 for a cylinder using Eq. (2.23).

The local forms of our curved spacetime postulates show that in many re-

spects a small region of a curved spacetime is like a small region of a ﬂat space-

time. We might suppose that all diﬀerences between the regions vanish as the

regions become smaller. This is not so.

To see this, refer to Fig. 2.7. Let ∆r be the small distance between objects

at the top and bottom of the cabin and let ∆a be the small tidal acceleration

between them. In a ﬂat spacetime ∆a = 0 for inertial particles. In the cabin

∆a = 0 , but ∆a →0 as ∆r →0 . However, ∆a/∆r →/ 0, as ∆r →0 : from

Eq. (2.1), da/dr = 2κM/r

3

. In a ﬂat spacetime, da/dr = 0. Here is a diﬀerence

between regions of the spacetime in the cabin and a ﬂat spacetime which does

not vanish as the regions become smaller.

The metric and geodesic postulates describe the behavior of clocks, light,

and inertial particles in a curved (or ﬂat) spacetime. But to apply these pos-

tulates, we must know the metric of the spacetime. Our ﬁnal postulate for

general relativity, the ﬁeld equation, determines the metric. Loosely speaking,

the equation determines the “shape” of a spacetime, how it is “curved”.

45

2.6 The Field Equation

We constructed the metric in Sec. 2.4 using local inertial frames. There

is obviously a relationship between the motion of local inertial frames and the

distribution of mass in a curved spacetime. Thus, there is a relationship between

the metric of a spacetime and the distribution of mass in the spacetime. The

ﬁeld equation gives this relationship. Schematically it reads

¸

quantity determined

by metric

=

¸

quantity determined

by mass/energy

. (2.24)

To specify the two sides of this equation, we need several deﬁnitions. Deﬁne

the Ricci tensor

R

jk

= Γ

p

tk

Γ

t

jp

−Γ

p

tp

Γ

t

jk

+∂

k

Γ

p

jp

−∂

p

Γ

p

jk

. (2.25)

Don’t panic over this convoluted deﬁnition: You need not have a physical un-

derstanding of the Ricci tensor. And while the R

jk

are extremely tedious to

calculate by hand, computers can readily calculate them for us.

The Ricci tensor is entirely determined by the metric. We shall see that

it contains information about the curvature K of two dimensional surfaces in

four dimensional spacetime. Like K, it involves second derivatives of the g

jk

(because the Γ

i

jk

involve the ﬁrst derivatives).

As with the metric g , we will use R to designate the Ricci tensor as a single

object, existing independently of any coordinate system, but which in a given

coordinate system acquires components R

jk

.

Deﬁne the curvature scalar R = g

jk

R

jk

.

We can now specify the left side of the schematic ﬁeld equation Eq. (2.24):

the quantity determined by the metric is the Einstein tensor

G = R−

1

2

Rg.

The right side of the ﬁeld equation is given by the energy-momentum tensor

T. It represents the source of the gravitational ﬁeld in general relativity. All

forms of matter and energy, including electromagnetic ﬁelds, and also pressures,

contribute to T. But for our purposes it will be suﬃcient to consider only matter

of a special form, called dust. In dust, each inﬁnitesimal element of matter is

inertial and thus moves on a geodesic. Gas in interstellar space which is thin

enough so that particle collisions are infrequent is dust.

We now deﬁne T for dust. Consider an event E with coordinates (y

i

). If

there is an element of dust at E, let (y

i

+ dy

i

) be a neighboring event on its

worldline. Let ds be the time between the events measured by a clock moving

with the element. Let ρ be the density of matter at E as measured by an observer

moving with the matter. Deﬁne

T

jk

= ρ

dy

j

ds

dy

k

ds

. (2.26)

We can now state our ﬁnal postulate for general relativity:

46

2.6 The Field Equation

The ﬁeld equation

G = −8πκT. (2.27)

Here κ is the Newtonian gravitational constant of Eq. (2.1).

The ﬁeld equation is the centerpiece of Einstein’s theory. It relates, event by

event, the curvature of spacetime, represented by G, to the density of matter

and energy in the spacetime, represented by T. The equation is a system of

second order nonlinear partial diﬀerential equations with the g

ij

as unknowns.

The metric, the geodesic equations, and the ﬁeld equation are applicable

in all coordinate systems. They are examples of our ability to describe curved

spacetimes intrinsically, i.e., without reference to a higher dimensional ﬂat space

in which they are curved. You should be impressed with the power of this

mathematics!

If ρ = 0 at some event, then the ﬁeld equation is R

jk

−

1

2

Rg

jk

= 0. Multiply

this by g

jk

:

0 = g

jk

R

jk

−

1

2

Rg

jk

= g

jk

R

jk

−

1

2

Rg

jk

g

jk

= R −

1

2

R4 = −R.

Substitute R = 0 into the ﬁeld equation to obtain

R = 0. (2.28)

At events in a spacetime where there is no matter we may use this vacuum ﬁeld

equation.

Before we can use the ﬁeld equation we must deal with a technical matter.

The indices on the metric and the Ricci tensor are subscripts: g

jk

and R

jk

. The

indices on the energy-momentum tensor are superscripts: T

jk

. By convention,

the placement of indices indicates how components transform under a change

of coordinates.

1

Subscripted components a

jk

transform covariantly:

¯ a

pq

= a

jk

∂y

j

∂¯ y

p

∂y

k

∂¯ y

q

. (2.29)

According to Exercise 2.4, the g

jk

transform covariantly. The R

jk

also transform

covariantly. (Do not attempt to verify this at home!) The Ricci scalar is the

same in all coordinate systems. Thus the left side of the ﬁeld equation, G

jk

=

R

jk

−

1

2

Rg

jk

, transforms covariantly.

Superscripted components a

jk

transform contravariantly:

¯ a

pq

= a

jk

∂¯ y

p

∂y

j

∂¯ y

q

∂y

k

. (2.30)

1

This is a convention of tensor algebra. Tensor algebra and calculus are powerful tools for

computations in general relativity. But we do not need them for a conceptual understanding

of the theory.

47

2.6 The Field Equation

Exercise 2.16. Show that the T

jk

transform contravariantly.

Since the G

jk

and T

jk

transform diﬀerently, we cannot take G

jk

= −8πκT

jk

as the components of the ﬁeld equation: even if this equation were true in one

coordinate system, it need not be in another. The solution is to raise the

covariant indices to contravariant indices: G

mn

= g

mj

g

nk

G

jk

.

Exercise 2.17. Show that if the G

jk

transform covariantly, then the G

mn

transform contravariantly.

Then we can take the components of the ﬁeld equation to be G

jk

= −8πκT

jk

.

If this equation is valid in any one coordinate system, then, since the two sides

transform the same between coordinate systems, it is true in all.

(Alternatively, we can lower the contravariant indices to covariant indices:

T

mn

= g

mj

g

nk

T

jk

, and use the equivalent equation G

jk

= −8πκT

jk

.)

The 00 component of the ﬁeld equation gives a simple and elegant relation-

ship between the curvatures K of certain surfaces through an event and the

density ρ at the event. To obtain it we must use Fermi normal coordinates.

The metric f of a local inertial frame at E satisﬁes Eq. (2.14), (f

mn

(E)) = f

◦

.

The metric of a geodesic coordinate system satisﬁes in addition Eq. (2.19),

∂

i

f

mn

(E) = 0. And in a spacetime, the metric of a Fermi normal coordinate

system satisﬁes further ∂

0

∂

i

f

mn

(E) = 0. We do not give a construction of these

coordinates.

Consider the element of matter at an event E. According to the local intertial

frame postulate, the element is instantaneously at rest in some local intertial

frame at E, which we take to have Fermi normal coordinates. Let K

12

be

the curvature of the surface formed by holding the time coordinate x

0

and the

spatial coordinate x

3

ﬁxed, while varying the other two spatial coordinates, x

1

and x

2

. Then G

00

= −(K

12

+K

23

+K

31

). Thus from the ﬁeld equation

K

12

+K

23

+K

31

= 8πκρ.

In particular, if ρ = 0, then K

12

+K

23

+K

31

= 0.

We close this chapter with a plausibility argument, based on reasonable

assumptions, which leads from the schematic ﬁeld equation Eq. (2.24) to the

ﬁeld equation Eq. (2.27). This will by no means be a proof of the equation,

but it should be convincing enough to make us anxious to confront the theory

with experiment in the next chapter. The reader may turn to the next chapter

without loss of continuity.

We take Eq. (2.24) to be a local equation, i.e., it will equate the two quan-

tities event by event.

Consider ﬁrst the right side of Eq. (2.24). It is reasonable to involve the

density of matter. In special relativity there are two eﬀects aﬀecting the density

of moving matter. First, the mass of a body moving with speed v increases by

a factor (1 − v

2

)

−

1

2

. According to Eq. (1.12), this factor is dx

0

/ds. Second,

from Appendix 4, an inertial body contracts in its direction of motion by the

same factor and does not contract in directions perpendicular to its direction of

48

2.6 The Field Equation

motion. Thus the density of moving matter is

T

00

= ρ

dx

0

ds

dx

0

ds

, (2.31)

where ρ is the density measured by an observer moving with the matter. As em-

phasized in Chapter 1, the coordinate diﬀerence dx

0

has no physical signiﬁcance

in and of itself. Thus Eq. (2.31) is only one component of a whole:

T

jk

= ρ

dx

j

ds

dx

k

ds

. (2.32)

This quantity represents matter in special relativity. A local inertial frame is

in many respects like an inertial frame in special relativity. Thus at the origin

of a local inertial frame we replace the right side of Eq. (2.24) with Eq. (2.32).

Transforming to global coordinates, the right side of Eq. (2.24) is the energy-

momentum tensor T.

Since the T

jk

transform contravariantly, the left side of Eq. (2.24) must

transform contravariantly. We denote it G

jk

in anticipation that it will turn

out to be the Einstein tensor. Thus the ﬁeld equation is of the form

G

jk

= −8πκT

jk

. (2.33)

(The factor −8πκ was inserted for later convenience.)

We shall make four assumptions which will uniquely determine the G

jk

.

(i) From Eq. (2.24), G depends on g. Assume that the left side of Eq.

(2.24), like the curvature K of a surface, depends only on the g

jk

and their ﬁrst

and second derivatives.

(ii) In special relativity ∂T

jk

/∂x

j

= 0. (For k = 0 this expresses conservation

of mass and for k = 1, 2, 3 it expresses conservation of the k component of

momentum.) Assume that ∂T

jk

/∂x

j

= 0 at the origin of local inertial frames.

According to a mathematical theorem of Lovelock, our assumptions already

imply that there is no loss of generality in taking the ﬁeld equation to be of the

form

A

R−

1

2

gR

+ Λg = −8πκT. (2.34)

Here A and Λ are constants.

(iii) Assume that a spacetime without matter is ﬂat. (We will revisit this

assumption in Sec. 4.4.) In a ﬂat spacetime the Christoﬀel symbols Eq. (2.17),

the Ricci tensor Eq. (2.25), and the curvature scalar R all vanish. Thus Eq.

(2.34) becomes Λg = 0. Thus Λ = 0.

(iv) Assume that Einstein’s theory agrees with Newton’s when Newton’s is

accurate, namely, for weak gravity with small velocities. We will show at the

end of Sec. 4.5 that this requires that A = 1. Thus Eq. (2.34) is precisely the

ﬁeld equation Eq. (2.27).

49

Chapter 3

Spherically Symmetric

Spacetimes

3.1 Stellar Evolution

Several applications of general relativity described in this chapter involve obser-

vations of stars at various stages of their life. We thus begin with a brief survey

of the relevant aspects of stellar evolution.

Clouds of interstellar gas and dust are a major component of our Milky Way

galaxy. Suppose some perturbing force causes a cloud to begin to contract by

self gravitation. If the mass of the cloud is small enough, then gas pressures

and/or mechanical forces will stop the contraction and a planet sized object

will result. For a larger cloud, these forces cannot stop the contraction. The

cloud will continue to contract and become hotter until thermonuclear reactions

begin. The heat from these reactions will increase the pressure in the cloud and

stop the contraction. A star is born! The star will continue to shine for millions

or billions of years until its nuclear fuel runs out and it begins to cool. Then

the contraction will begin again. The subsequent evolution of the star depends

on its mass.

For a star of ∼1.4 solar masses or less, the contraction will be stopped by

a phenomenon known as degenerate electron pressure - but not until enormous

densities are reached. For example, the radius of the Sun will decrease by a

factor of 10

2

- to about 1000 km - and its density will thus increase by a factor

of 10

6

, to about 10

6

g/cm

3

! The star is a white dwarf . They are common. For

example, Sirius, the brightest star in the sky, is part of a double star system.

Its dim companion is a white dwarf.

If the white dwarf is a member of a binary star system, it sometimes accretes

matter from its companion. If it accretes suﬃcient matter, it will erupt in a

runaway thermonuclear explosion, called a Type Ia supernova, destroying the

star. At its brightest, a Type Ia supernova has a luminosity over a billion times

that of the Sun.

50

3.1 Stellar Evolution

For a star over 1.4 solar masses, degenerate electron pressure cannot stop the

contraction. Eventually a supernova of another kind, called Type II, occurs. The

outer portions of the star blow into interstellar space. A supernova was recorded

in China in 1054. It was visible during the day for 23 days and outshone all

other stars in the sky for several weeks. Today we see the material blown from

the star as the Crab Nebula.

One possible result of a Type II supernova is a neutron star. In a neutron

star, fantastic pressures force most of the electrons to combine with protons to

form neutrons. Degenerate neutron pressure prevents the star from collapsing

further. A typical neutron star has a radius of 10 km and a density of 10

14

g/cm

3

!

Neutron stars have manifested themselves in two ways. They often spin rapidly

- up to nearly 1000 times a second. For poorly understood reasons, there can

be a small region of the star that emits radio and optical frequency radiation in

a rather narrow cone. If the Earth should happen to be in the direction of the

cone periodically as the star rotates, then the star will appear to pulse at the

frequency of rotation of the star. The neutron star is a pulsar. The ﬁrst pulsar

was discovered in 1968; many are now known. There is a pulsar at the center

of the crab nebula.

There are also neutron stars which pulse strongly in X-rays. They are always

members of a binary star system. We discuss these systems in Sec. 3.6.

Even degenerate neutron pressure cannot stop the contraction if the star is

too massive. A massive star can collapse to a black hole. We discuss black holes

in Sec. 3.6.

We close our catalog of remarkable astronomical objects with quasars, dis-

covered in 1963. Quasars sit at the center of some galaxies, mostly distant. A

typical quasar emits 100 times the energy of our entire Milky Way galaxy from

a region 10

17

times smaller!

51

3.2 The Schwartzschild Metric

3.2 The Schwartzschild Metric

In this section we obtain the metric for the spacetime around a spherically

symmetric object such as the Sun. This will enable us in the next section to

compare the predictions of general reltivity with observations of the motion of

planets and light in our solar system.

Exercise 3.1. Show that in spherical coordinates the ﬂat spacetime metric

Eq. (1.14) becomes

ds

2

= dt

2

−dr

2

−r

2

dΩ

2

, (3.1)

where from Eq. (2.11), dΩ

2

= dφ

2

+ sin

2

φ dθ

2

is the metric of the unit sphere.

Hint: Think geometrically.

How can Eq. (3.1) change in a curved spherically symmetric spacetime? In

such a spacetime:

• The θ and φ coordinates can have their usual meaning. (As in a curved

surface, angles can be measured in the usual way in a curved spacetime.)

• dθ and −dθ produce the same ds, as do dφ and −dφ. Thus none of the

terms drdθ, drdφ, dθdφ, dtdθ, or dtdφ can appear in the metric.

• The surface t = t

o

, r = r

o

has the metric of a sphere, although not

necessarily of radius r

o

.

Thus the metric is of the form

ds

2

= e

2µ

dt

2

−2udtdr −e

2ν

dr

2

−r

2

e

2λ

d Ω

2

, (3.2)

where µ, ν, u, and λ are unknown functions of t and r, but not, by symmetry, of

θ or φ. We write some of the coeﬃcients as exponentials for convenience.

The coordinate change ¯ r = re

λ

eliminates the e

2λ

factor. (Then rename the

radial coordinate r.) A change in the t coordinate eliminates the dtdr term. See

Appendix 12. These coordinate changes put the metric in a simpler form:

ds

2

= e

2µ

dt

2

−e

2ν

dr

2

−r

2

d Ω

2

. (3.3)

This is as far as we can go with symmetry and coordinate changes. We must

use the ﬁeld equation to determine µ and ν. Since we are interested only in

the spacetime outside the central object, we use the vacuum ﬁeld equation Eq.

(2.28): R = 0. The components of R are given by Eq. (2.25). They are best

calculated with a computer.

We ﬁrst set R

tr

=−2 (∂ν/∂t)/r = 0 . Thus ∂ν/∂t = 0, i.e., ν depends only

on r. Now

R

tt

=

µ

−µ

ν

+µ

2

+ 2 µ

r

−1

R

rr

=

µ

−µ

ν

+µ

2

−2 ν

r

−1

e

−4ν

(3.4)

R

φφ

=

−rµ

+rν

+ e

2ν

−1

e

−2ν

,

52

3.2 The Schwartzschild Metric

where primes indicate diﬀerentiation with respect to r.

Set R

tt

= 0 and R

rr

= 0, cancel the exponential factor, and subtract:

µ

+ ν

= 0. Now R

φφ

= 0 implies 2 rν

+ e

2ν

− 1 = 0, or (r e

−2ν

)

= 1.

Integrate: e

−2ν

= 1 − 2 m/r, where 2 m is a constant of integration. Since ν

does not depend on t, neither does m. The metric Eq. (3.3) is now of the form

ds

2

=

1 −

2m

r

e

2(µ+ν)

dt

2

−

1 −

2m

r

−1

dr

2

−r

2

d Ω

2

.

Since µ

+ ν

**= 0, µ + ν is a function only of t. Thus the substitution
**

d

¯

t = e

µ+ν

dt eliminates the e

2(µ+ν)

factor. We arrive at our solution, the

Schwartzschild metric, obtained by Karl Schwartzschild in 1916:

Schwartzschild Metric

ds

2

=

1 −

2m

r

dt

2

−

1 −

2m

r

−1

dr

2

−r

2

d Ω

2

. (3.5)

We shall see later that

m = κM, (3.6)

where κ is the Newtonian gravitational constant of Eq. (2.1) and M is the mass

of the central object. For the Sun,

m = 4.92 10

−6

sec = 1.47 km.

If r 2m, then the Schwartzschild metric Eq. (3.5) and the ﬂat spacetime

metric Eq. (3.1) are nearly identical. Thus if r 2m, then for most purposes r

may be considered radial distance and t time measured by slowly moving clocks.

Exercise 3.2. Show that the time ds measured by a clock at rest at r is

related to the coordinate time dt by

ds =

1 −

2m

r

1

2

dt. (3.7)

Exercise 3.3. Show that the Schwartzschild metric predicts the sum of

the results of Exercises 1.9 and 2.1 for the diﬀerence between the clocks in the

Hafele-Keating experiment: ∆s

a

− ∆s

g

=

1

2

(v

2

a

−v

2

g

) +gh

**∆t . Thus general
**

relativity gives both the time dilation and gravitational redshift diﬀerences, and

nothing more.

It is interesting to note that our derivation of the Schwartzschild metric

applies to the interior of a hollow sphere surrounded by a spherically symmetric

distribution of matter. In this case m = 0; otherwise there would be a singularity

in the metric at r = 0. Thus the Schwartzschild metric reduces to the ﬂat

spacetime metric; the matter outside the sphere exerts no gravitational inﬂuence

inside the sphere.

It is remarkable that the Schwartzschild metric is independent of t even

though we did not assume this. Thus if the central object is expanding or

53

3.2 The Schwartzschild Metric

contracting in a spherically symmetric way, then the metric cannot change.

Such an object cannot emit gravitational radiation – propagated disturbances

in the metric caused by matter in motion.

The mathematical description of gravitational radiation and the physical

interpretation of the mathematical description are complicated matters beyond

the scope of this book. Violent astronomical phenomena, such as supernovae

(see Sec. 3.1), collisions of stars (see Sec. 3.5), and the big bang (see Sec. 4.1)

emit gravitational radiation. The radiation travels at the speed of light. It

can carry energy away from its source. Several detectors have been built and

are being built to try to detect gravitational radiation. So far none has been

detected. (But see Sec. 3.5.) The diﬃculty in detecting this radiation is its

extreme weakness.

We now investigate the gravitational redshift using the Schwartzschild met-

ric. Emit pulses of light radially outward from (t

e

, r

e

) and (t

e

+∆t

e

, r

e

). Observe

them at (t

o

, r

o

) and (t

o

+ ∆t

o

, r

o

). Since the Schwartzschild metric is time in-

dependent, the worldline of the second pulse is simply a tranlation of the ﬁrst

by ∆t

e

. Thus ∆t

o

= ∆t

e

. Let ∆s

e

be the time between the emission events as

measured by a clock at r

e

, with ∆s

o

deﬁned similarly. By Eqs. (1.6) and Eq.

(3.7) an observer at r

o

ﬁnds a redshift

z =

∆s

o

∆s

e

−1 =

1 −2m/r

o

1 −2m/r

e

−1. (3.8)

Exercise 3.4. Show that if r

e

2m and r

o

− r

e

is small, then Eq. (3.8)

reduces to Eq. (2.2), the gravitational redshift formula derived in connection

with the terrestrial redshift experiment. Note that c = 1 in Eq. (3.8).

For light emitted at the surface of the Sun and received at Earth, Eq. (3.8)

gives z = 2 10

−6

. This is diﬃcult to measure but it has been veriﬁed within

7%. Light from a star with the mass of the Sun but a smaller radius will, by Eq.

(3.8), have a larger redshift. For example, light from the white dwarf companion

to Sirius has a gravitational redshift z = 3 10

−4

. And a gravitational redshift

z = .35 has been measured in X-rays emitted from the surface of a neutron star.

Exercise 3.5. Show that a clock on the Sun will lose 63 sec/year compared

to a clock far from the Sun. (The corresponding losses for the white dwarf and

neutron star just mentioned are 2.6 hours and 95 days, respectively.)

An accurate measurement of the gravitational redshift was made in 1976 by

an atomic clock in a rocket. The reading of the clock was compared, via radio,

with one on the ground during the two hour ﬂight of the rocket. Of course

the gravitational redshift changed with the changing height of the rocket. After

taking into account the redshift due to the motion of the rocket, the gravitational

redshift predicted by general relativity was conﬁrmed within 7 parts in 10

5

.

54

3.2 The Schwartzschild Metric

To determine the motion of inertial particles and light in a Schwartzschild

spacetime, we must solve the geodesic equations Eq. (2.21). They are:

¨

t +

2m/r

2

1 −2m/r

˙ r

˙

t = 0 (3.9)

¨ r +

m(1 −2m/r)

r

2

˙

t

2

−

m/r

2

1 −2m/r

˙ r

2

−r(1 −2m/r)

˙

φ

2

+ sin

2

φ

˙

θ

2

= 0 (3.10)

¨

θ +

2

r

˙ r

˙

θ + 2 cotφ

˙

φ

˙

θ = 0 (3.11)

¨

φ +

2

r

˙ r

˙

φ −sin φ cos φ

˙

θ

2

= 0. (3.12)

From spherical symmetry a geodesic lies in a plane. Let it be the plane

φ = π/2 . (3.13)

This is a solution to Eq. (3.12).

Exercise 3.6. Use Eqs. (3.10) and (3.5) to show that ds = (1 −3m/r)

1

2

dt

for a circular orbit at φ = π/2. (Cf. Exercise 3.2.) This shows that a clock

(with ds > 0) can have a circular orbit only for r > 3m. It also shows that light

(with ds = 0) can orbit at r = 3m.

The circular orbits of the exercise are of course possible only if the central

object is inside the sphere r = 3m. This is not so for ordinary objects, but it is

for a black hole. See Sec. 3.6.

Integrate Eqs. (3.9) and (3.11):

˙

t (1 −2m/r) = B (3.14)

˙

θ r

2

= A, (3.15)

where A and B are constants of integration. Physically, A and B are conserved

quantities: A is the angular momentum per unit mass along the geodesic and

B is the energy per unit mass.

Exercise 3.7. Diﬀerentiate Eqs. (3.14) and (3.15) to obtain Eqs. (3.9)

and (3.11). (Remember that φ = π/2 .)

Substitute Eqs. (3.13)-(3.15) into Eq. (3.10) and integrate:

˙ r

2

1 −2m/r

+

A

2

r

2

−

B

2

1 −2m/r

= −E =

0 for light

−1 for inertial particles,

(3.16)

where E is a constant of integration. The values given for E are veriﬁed by

substituting Eqs. (3.13)-(3.16) into the Schwartzschild metric Eq. (3.5), giving

(ds/dp)

2

= E. For light ds = 0 (see Eq. (2.16)) and so E = 0. For an inertial

particle we may take p = s (see Eq. (2.21)) and so E = 1.

55

3.2 The Schwartzschild Metric

The partially integrated geodesic equations Eqs. (3.13)-(3.16) will be the

starting point for the study of the motion of planets and light in the solar system

in the next section. We took several side journeys enroute to the equations, so

it might be useful to outline the steps we used to obtain them:

• We obtained a general form for the metric of a spherically symmetric

spacetime, Eq. (3.2).

• The Ricci tensor involves second derivatives of a metric. In Eqs. (3.4) we

wrote some of the components of this tensor for the spherically symmetric

metric.

• We wrote the vacuum ﬁeld equations for the metric by setting the com-

ponents of the Ricci tensor to zero.

• We solved the vacuum ﬁeld equations to obtain the Schwartzschild metric,

Eq. (3.5).

• The geodesic equations involve ﬁrst derivatives of a metric. In Eqs. (3.9)-

(3.12) we wrote the geodesic equations for the Schwartzschild metric. They

are of second order.

• We integrated the second order geodesic equations to obtain ﬁrst order

Eqs. (3.13)-(3.16).

We are now able to evaluate the constant m in the Schwartzschild metric

Eq. (3.5). Consider radial motion of an inertial particle. By Eq. (3.15), A = 0.

Thus Eq. (3.16) becomes

dr

ds

= ±

B

2

−1 +

2m

r

1

2

, (3.17)

the sign chosen according as the motion is outward or inward. Diﬀerentiate Eq.

(3.17) and substitute Eq. (3.17) into the result:

d

2

r

ds

2

= −

m

r

2

. (3.18)

For a distant slowly moving particle, ds ≈ dt. Since the Newtonian theory

applies to this situation, Eqs. (2.1) and (3.18) must coincide. Thus m = κM,

in agreement with Eq. (3.6).

56

3.3 The Solar System Tests

3.3 The Solar System Tests

In this section we compare the predictions of general relativity with observations

of the motion of planets and light in our solar system. We discuss three general

relativistic eﬀects: perihelion advance, light deﬂection, and light retardation.

Perihelion Advance. We ﬁrst solve the geodesic equations for the planets.

Set u = 1/r. By Eq. (3.15),

˙ r =

dr

du

du

dθ

˙

θ = −u

−2

du

dθ

Ar

−2

= −A

du

dθ

.

Substitute this into Eq. (3.16), multiply by 1 −2mu, diﬀerentiate with respect

to θ, and divide by 2A

2

du/dθ:

d

2

u

dθ

2

+u = mA

−2

E + 3mu

2

. (3.19)

For an inertial particle E = 1, and Eq. (3.19) is identical to the equation of

motion Eq. (A.5) derived from the Newtonian theory except for the term 3mu

2

.

For the planets this term is very small. Thus, although Einstein’s theory is con-

ceptually entirely diﬀerent from Newton’s, it gives nearly the same predictions

for the planets. This is necessary for any theory of gravity, as Newton’s theory

is very accurate for the planets.

Exercise 3.8. Among the planets, Mercury has the largest ratio of the

terms on the right side of Eq. (3.19). Show that it is less than 10

−7

.

As a ﬁrst approximation, solve Eq. (3.19) without the small 3mu

2

term:

u = mA

−2

[1 +e cos(θ −θ

p

)] . (3.20)

This is the equation of an ellipse with eccentricity e and perihelion (point of

closest approach to the Sun) at θ = θ

p

. Set θ = θ

p

in Eq. (3.20):

mA

−2

= r

−1

p

(1 +e)

−1

(3.21)

To obtain a more accurate solution of Eq. (3.19), substitute Eqs. (3.20) and

(3.21) into the right side of Eq. (3.19) and solve:

u = r

−1

p

(1 +e)

−1

1 +e

cos(θ −θ

p

) + 3mr

−1

p

(1 +e)

−1

θ sin(θ −θ

p

)

¸

+ mr

−2

p

(1 +e)

−2

¸

3 +

1

2

e

2

[ 3 −cos(2(θ −θ

p

))]

¸

. (3.22)

Drop the last term, which stays small because of the factor mr

−2

p

. For small α,

cos β +αsin β ≈ cos β cos α + sin β sin α = cos(β −α).

Use this approximation in Eq. (3.22):

u = r

−1

p

(1 +e)

−1

1 +e cos

¸

θ −

θ

p

+

3mθ

r

p

(1 +e)

. (3.23)

57

3.3 The Solar System Tests

Since 3m/[ r

p

(1 +e) ] is small, this may be considered the equation of an ellipse

with perihelion advanced by 6mπ/[ r

p

(1 +e) ]/revolution.

Exercise 3.9. Show that for Mercury the predicted perihelion advance is

43.0 arcsecond/century.

The perihelion of Mercury is observed to advance ∼500 arcsec/century. Since

1859 astronomers explained all but about 43 arcsec as due to slight gravitational

eﬀects of other planets. The 43 arcsec was the only known discrepancy between

Newton’s theory and observation when Einstein published the general theory

of relativity. Less than an arcminute per century! This is the remarkable

accuracy of Newton’s theory. The explanation of the discrepancy was the ﬁrst

observational veriﬁcation of general relativity. Today this prediction of general

relativity is veriﬁed within .5%.

Light Deﬂection. Ac-

Fig. 3.1: Deﬂection of light by the sun.

cording to general relativity,

light passing near the Sun

will be deﬂected. See Fig.

3.1. Consider light which

grazes the Sun at θ = 0 when

u = u

p

= 1/r

p

. Set E = 0 in

Eq. (3.19) for light to give

the geodesic equation

d

2

u

dθ

2

+u = 3mu

2

. (3.24)

Ignore the small 3mu

2

term to obtain a straight line as an approximate solution:

u = u

p

cos θ. Substitute this into the right side of Eq. (3.24) and solve to obtain

a better approximation:

u = u

p

cos θ +

1

2

mu

2

p

(3 −cos 2θ).

Set u = 0 (r = ∞; θ = ±(π/2 +δ/2) and approximate

cos(π/2 +δ/2) = −sin δ/2 ≈ −δ/2, cos(π +δ) = −cos δ ≈ −1

to obtain the observed deﬂection angle δ = 4m/r

p

. (The Newtonian equation

Eq. (2.1) predicts half of this deﬂection.)

Exercise 3.10. Show that for the Sun, δ = 1.75 arcsec.

Stars can be seen near the Sun only during a solar eclipse. Sir Arthur Stanley

Eddington organized an expedition to try to detect the deﬂection during the

eclipse of 1919. He conﬁrmed the prediction of general relativity within about

20%. Later eclipse observations have improved the accuracy, but not by much.

In 1995 the predicted deﬂection was veriﬁed within .01% using radio waves

emitted by a quasar. Radio waves have two advantages. First, several radio

telescopes working together can measure angles more accurately than optical

58

3.3 The Solar System Tests

telescopes. Second, an eclipse is not necessary, as radio sources can be detected

during the day. The accuracy of such measurements is now better than 10

−4

arcsec. Light arriving from a quasar at an angle of 90

◦

from the sun is deﬂected

by 4 10

−3

arcsec by the sun, and so the deﬂection can be measured.

Fig. 3.2: Gravitational lens.

A spectacular example of the gravita-

tional deﬂection of light was discovered in

1979. Two quasars, 6 arcsec apart, are in

fact the same quasar! Fig. 3.2 shows the

cause of the double image of the quasar.

The galaxy is a gravitational lens. Many

multiple (up to 7) image quasars are now

known. It is not necessary that the quasar

be exactly centered behind the galaxy for

this to occur. But in many cases this is

nearly so; then the quasar appears as one

or several arcs around the galaxy. And in at least one case the quasar appears

as a ring around the galaxy! A gravitational lens can also brighten a distant

quasar by up to 100 times, enabling astronomers to study them better.

Fig. 3.3: Radar echo delay.

Light Retardation. Radar can be

sent from Earth, reﬂected oﬀ a planet, and

detected upon its return to Earth. If this

is done as the planet is about to pass be-

hind the Sun, then according to general

relativity, the radar is slowed by the Sun’s

gravity. See Fig. 3.3.

To calculate the time t for light to go

from Earth to Sun, use Eq. (3.14):

dr

dp

=

dr

dt

dt

dp

=

dr

dt

B

1 −

2m

r

−1

.

Set r = r

p

and ˙ r[

r=rp

= 0 in Eq. (3.16):

A

2

B

−2

= r

2

p

1 −

2m

r

p

−1

.

Divide Eq. (3.16) by B

2

, substitute the above two equations into the result,

separate the variables, and integrate:

t =

re

rp

¸

1 −

2m

r

−2

1 −

1−2m/r

1−2m/rp

rp

r

2

¸

1

2

dr. (3.25)

In Appendix 13 we approximate the integral:

t = (r

2

e

−r

2

p

)

1

2

+ 2mln(2r

e

/r

p

) +m. (3.26)

59

3.3 The Solar System Tests

The ﬁrst term, (r

2

e

− r

2

p

)

1

2

, is, by the Pythagorean theorem, the time required

for light to travel in a straight line in a ﬂat spacetime from Earth to Sun. The

other terms represent a lengthening of this ﬂat spacetime t.

Add analogous delay terms for the path from Sun to Planet, double to include

the return trip, and ﬁnd for Mercury a delay of 2.4 10

−4

sec. One diﬃculty

in performing the experiment lies in determining the positions of the planets

in terms of the r coordinate of the Schwartzschild metric accurately enough to

calculate the ﬂat spacetime time. This diﬃculty and others have been overcome

and the prediction veriﬁed within 5%. Signals from the Viking spacecraft on

Mars have conﬁrmed this eﬀect within .1%

Quasars can vary in brightness on a timescale of months. We see variation in

the southern image of the original double quasar about 1

1

2

years after the same

variation of the northern image. The southern image of the quasar is 1 arcsec

from the lensing galaxy; that of the northern is 5 arcsec. Thus the light of the

southern image travels less distance than that of the southern, causing a delay in

the northern image. But this is more than compensated for by the gravitational

delay of the light of the southern image passing closer to the lensing galaxy.

Another eﬀect of general relativity, Willem deSitter’s geodetic eﬀect, is mo-

tivated in Figs. 3.4 and 3.5. In Fig. 3.4, a vector in a plane is moved parallel to

itself from A around a closed curve made up of geodesics. The vector returns to

A with its original direction. In Fig. 3.5 a vector on a sphere is moved parallel

to itself (i.e., parallel to itself in local planar frames), from A around a closed

curve made of geodesics. The vector returns to A rotated through the angle α.

This is a manifestation of curvature.

Fig. 3.4: A parallel transported vector

returns to A with its original direction.

Fig. 3.5: A parallel transported vector

returns to A rotated through angle α.

The axis of rotation of a gyroscope moves parallel to itself in inertial frames in

a ﬂat spacetime . The same is true in local inertial frames in a curved spacetime.

But over a worldline the orientation of the axis with respect to the distant stars

can change. This is the geodetic eﬀect. For example, a gyroscope in a circular

orbit at coordinate r in a Schwartzschild spacetime will precess about an axis

normal to the plane of the orbit by an angle 2π[1 −(1 −3m/r)

1

2

]/revolution in

the direction of the orbit. The Earth-Moon system is a “gyroscope” orbiting

the Sun. The geodetic eﬀect predicts a change of the axis of rotation of ∼.02

arcsec/yr. The lunar laser experiment has conﬁrmed this to 1%.

60

3.4 Kerr Spacetimes

3.4 Kerr Spacetimes

The Schwartzschild metric describes the spacetime around a spherically sym-

metric object. In particular, the object cannot be rotating. The Kerr metric,

discovered only in 1963 by Roy Kerr, describes the spacetime outside a rotating

object which is symmetric around its axis of rotation and unchanging in time:

Kerr Metric

ds

2

=

1 −

2mr

Σ

dt

2

+

4mar sin

2

φ

Σ

dt dθ −

Σ

∆

dr

2

−Σdφ

2

−

r

2

+a

2

+

2ma

2

r sin

2

φ

Σ

sin

2

φdθ

2

, (3.27)

where m = κM, as in the Schwartzschild metric; a = J/M, the angular momen-

tum per unit mass of the central object; Σ = r

2

+a

2

cos

2

φ; ∆ = r

2

−2mr +a

2

;

and the axis of rotation is the z-axis.

Exercise 3.11. Show that if the central object is not spinning, then the

Kerr metric Eq. (3.27) reduces to the Schwartzschild metric Eq. (3.5).

Note the surfaces t = const, r = const in a Kerr spacetime do not have the

metric of a sphere.

All of the gravitational eﬀects that we have discussed until now are eﬀects of

the mass of an object. Gravitomagnetism is a gravitational eﬀect of the motion

of an object. Its name refers to an analogy with electromagnetism: an electric

charge at rest creates an electric ﬁeld, while a moving charge also creates a

magnetic ﬁeld.

The Lense-Thirring or frame dragging gravitomagnetic eﬀect causes a change

in the orientation of the axis of rotation of a small spinning object in the space-

time around a large spinning object. An Earth satellite is, by virtue of its orbital

motion, a spinning object, whose axis of rotation is perpendicular to its orbital

plane. The Kerr metric predicts that the axis of a satellite in polar orbit at

two Earth radii will change by ∼.03 arcsec/yr. This is a shift of ∼1 m/yr in

the intersection of the orbital and equatorial planes. Observations of two LA-

GEOS satellites in non-polar orbits have probably detected the Lense-Thirring

eﬀect within 30% of the prediction of general relativity. This is the ﬁrst direct

measurement of gravitomagnetism.

Satellites in the same circular equatorial orbit in a Kerr spacetime but with

opposite directions take diﬀerent coordinate times ∆t

±

for one complete orbit

θ → θ ± 2π. We derive this gravitomagnetic clock eﬀect from the r-geodesic

equation for circular equatorial orbits in the Kerr metric:

˙

t

2

−2a

˙

t

˙

θ +

a

2

−

r

3

m

˙

θ

2

= 0. (3.28)

Solve for dt/dθ:

dt

dθ

±

= ±

r

3

m

1

2

+a, (3.29)

61

3.4 Kerr Spacetimes

where ± corresponds to the orbits θ →θ ±2π.

Integrate Eq. (3.29) over one orbit by multiplying the right side of the

equation by ±2π: ∆t

±

= 2π(r

3

/m)

1

2

±2πa. Then ∆t

+

−∆t

−

= 4πa; an orbit

in the sense of a takes longer than an opposite orbit.

Exercise 3.12. Show that for the Earth, ∆t

+

−∆t

−

= 2 10

−7

sec. This

has not been measured.

62

3.5 The Binary Pulsar

3.5 The Binary Pulsar

As we have seen, the diﬀerences between Einstein’s and Newton’s theories of

gravity are exceedingly small in the solar system. Large diﬀerences require

strong gravity and/or large velocities. Both are present in a remarkable binary

star system discovered in 1974. Both stars are 1.4 solar mass neutron stars.

Their orbital period is eight hours, and the two stars are about a solar diameter

apart. Imagine!

By great good fortune, one of the neutron stars is a pulsar spinning at 17

revolutions/sec. The pulsar’s pulses serve as the ticking of a very accurate clock

in the system. Several relativistic eﬀects have been measured by analyzing

the arrival times of the pulses at Earth. These measurements provide the most

probing tests of general relativity to date, earning the system the title “Nature’s

gift to relativists”.

The periastron advance of the pulsar is 4.2

◦

/year – 35,000 times that of

Mercury. (Perihelion refers to the Sun; periastron is the generic term.) The

Schwartzschild metric cannot predict this because neither star is an inertial

particle in the gravity of the other. More complicated techniques using the full

ﬁeld equation, rather than the vacuum ﬁeld equation, must be used.

Calculations predict that the orbital period of the system should decrease by

10

−4

sec/year due to energy loss from the system due to gravitational radiation.

Several changing relativistic eﬀects aﬀecting the arrival times at Earth of the

pulsar’s pulses must be taken into account while measuring a change in the

period: the gravitational redshift of the rate of clocks on Earth as the Earth’s

distance from the Sun changes over the course of the year, the gravitational

redshift of the time between the pulses as the pulsar moves in its highly elliptical

(eccentricity = .6) orbit, a “slowing” of the pulsar clock due to its motion, and

a delay in the pulses, analogous to the delay measured in the radar echo delay

experiment. By 1978 the period was known to decrease by about the amount

predicted by general relativity. The prediction is now conﬁrmed within .5%.

This is the best empirical evidence for the existence of gravitational radiation.

As the binary system loses energy by gravitational radiation, the stars come

closer together. They will collide in about 300 million years. The orbital period

just before the collision will be of the order of 10

−3

sec!

63

3.6 Black Holes

3.6 Black Holes

The Schwartzschild metric Eq. (3.5) has a singularity at the Schwartzschild

radius r = 2m: g

rr

= ∞ there. The r = 2m surface is called the (event)

horizon. For the Sun, the Schwartzschild radius is 3 km and for the Earth it

is .9 cm. This is well inside these bodies, and so the singularity has no eﬀect

on the Schwartzschild metric, a solution of the vacuum ﬁeld equation valid only

outside the central body. We may inquire however about the properties of an

object which is inside its Schwartzschild radius. The object is then a black hole.

Exercise 3.13 shows that the singularity at the Schwartzschild radius is only

a coordinate singularity; there is no singularity in the spacetime. An observer

crossing the horizon would not notice anything special. To see a coordinate

singularity in a more familiar setting, substitute u = 1/r in the ﬂat surface polar

coordinate metric ds

2

= dr

2

+ r

2

dθ

2

to obtain ds

2

= u

−4

du

2

+ u

−2

dθ

2

. There

a coordinate singularity at the origin, but there is no geometric singularity.

Exercise 3.13. Introduce Painlev`e coordinates with the coordinate change

dt = d

¯

t +

(2mr)

1

2

r −2m

dr

in the Schwartzschild metric (then drop the bar) to obtain the metric

ds

2

= dt

2

−

dr +

2m

r

1

2

dt

2

−r

2

d Ω

2

.

The only singularity is at r = 0, a true spacetime singularity.

Suppose a material particle or pulse of light moves outward (not necessarily

radially) from r

1

= 2m to r

2

. Since ds

2

≥ 0 and dΩ

2

≥ 0 , the Schwartzschild

metric Eq. (3.5) shows that dt ≥ (1 − 2m/r)

−1

dr. Thus the total coordinate

time ∆t satisﬁes

∆t ≥

r2

2m

1 −

2m

r

−1

dr =

r2

2m

1 +

2m

r −2m

dr = ∞;

neither matter nor light can escape a Schwartzschild black hole! The same

calculation shows that it takes an inﬁnite coordinate time ∆t for matter or light

to enter a black hole.

Exercise 3.14. Suppose an inertial object falling radially toward a black

hole is close to the Schwartzschild radius, so that r−2m is small. Use Eqs. (3.14)

and (3.17) and approximate to show that dr/dt = (2m−r)/2m. Integrate and

show that r −2m decreases exponentially in t.

64

3.6 Black Holes

According to Eq. (3.8), the redshift of an object approaching a

Schwartzschild black hole increases to inﬁnity as the object approaches the

Schwartzschild radius. A distant observer will see the rate of all physical pro-

cesses on the object slow to zero as it approaches the Schwartzschild radius. In

particular, its brightness (rate of emission of light) will dim to zero and it will

eﬀectively disappear.

On the other hand, an inertial object radially approaching a Schwartzschild

black hole will cross the Schwartzschild radius in a ﬁnite time according to a

clock it carries! To see this, observe that by the metric postulate Eq. (2.15),

dr/ds is the rate of change of r measured by a clock carried by the object.

By Eq. (3.17), [dr/ds[ increases as r decreases. Thus the object will cross the

Schwartzschild radius, never to return, in a ﬁnite proper time ∆s measured by

its clock.

Most stars rotate. Due to conservation of angular momentum, a rotating

star collapsed to a black hole will spin very rapidly. We must use the Kerr

metric to describe a rotating black hole.

For a Schwartzschild black hole, g

rr

=

Fig. 3.6: The horizon and static

limit of a Kerr black hole.

∞ and g

tt

= 0 on the surface r = 2m,

the horizon. We have seen that there is

no singularity in the spacetime there. For

a Kerr black hole, g

rr

= ∞ on the surface

∆ = 0, and g

tt

= 0 on the surface Σ = 2mr.

Fig. 3.6 shows the axis of rotation of the

Kerr black hole, the inner surface ∆ = 0

(called the horizon), and the outer surface

Σ = 2mr (called the static limit). There

is no singularity in the spacetime on either

surface.

Exercise 3.15. a. Show that the horizon is at r = m+ (m

2

−a

2

)

1

2

.

b. Show that the static limit has equation r = m+ (m

2

−a

2

cos

2

φ)

1

2

.

As with the Schwartzschild horizon, it is impossible to escape the Kerr hori-

zon, even though it is possible to reach it in ﬁnite proper time. The proofs are

more diﬃcult than for the Schwartzschild metric. They are not given here.

The crescent shaped region between the horizon and the static limit is the

ergosphere. In the ergosphere, the dt

2

, dr

2

, dφ

2

, and dθ

2

terms in the Kerr

metric Eq. (3.27) are negative. Since ds

2

≥ 0 for both particles and light, the

dtdθ term must be positive, i.e., a(dθ/dt) > 0. In the ergosphere it is impossible

to be at rest, and motion must be in the direction of the rotation!

65

3.6 Black Holes

General relativity allows black holes, but do they exist in Nature? The

answer is almost certainly “yes”. At least two varieties are known: solar mass

black holes, of a few solar masses, and supermassive black holes, of 10

6

− 10

10

solar masses.

We saw in Sec. 3.1 that the Earth, the Sun, white dwarf stars, and neutron

stars are stabilized by diﬀerent forces. No known force can stabilize a supernova

remnant larger than ∼3 solar masses; the remnant will collapse to a black hole.

Since black holes emit no radiation, they must be observed indirectly. Several

black holes of a few solar masses are known in our galaxy. They are members

of X-ray emitting binary star systems in which a normal star is orbiting with

a compact companion. Orbital data show that in some of these systems, the

companion’s mass is small enough to be a neutron star. But in others, the

companion is well over the three solar mass limit for neutron stars, and is

therefore believed to be a black hole.

During “normal” periods, gas pulled from the normal star and heated vi-

olently on its way to the compact companion causes the X-rays. Some of the

systems with a neutron star companion exhibit periods of much stronger X-ray

emission. Sometimes the cause is gas suddenly crashing onto the surface of the

neutron star. Sometimes the cause is a thermonuclear explosion of material

accumulated on the surface. Systems with companions over three solar masses

never exhibit such periods. Presumably this is because the companions are black

holes, which do not have a surface.

Many, if not most, galaxies have a supermassive black hole at their center.

X-ray observations of several galactic nuclei indicate that matter orbiting the

nuclei at 3-10 Schwartzschild radii at 10% of the speed of light. Quasars are

apparently powered by a massive black hole at the center of a galaxy.

A strong radio source at the center of our Milky Way galaxy has a radius of

about one light hour. There is a star orbiting the source with eccentricity .87,

closest approach to the source 17 light days, and with a period of 15 years. This

implies that the source is 3.7 million solar masses. A black hole at the center

seems to be the only possibility.

66

Chapter 4

Cosmological Spacetimes

4.1 Our Universe I

This chapter is devoted to cosmology, the study of the universe as a whole. Since

general relativity is our best theory of space and time, we use it to construct a

model of the spacetime of the entire universe. This is an audacious move, but

the model is very successful, as we shall see.

We begin with a description of the universe as seen from Earth. The Sun

is but one star in 200 billion or so bound together gravitationally to form our

galaxy, the Milky Way. The Milky Way is about 100,000 light years across. It

is but one galaxy in tens of billions in the visible universe. The nearest major

galaxies are a few million light years away. Galaxies are inertial objects.

Fig. 4.1: Balloon analogy to

the universe.

But they are not at rest. All galaxies, except a

few nearby, recede from us with a velocity v pro-

portional to their distance d from us:

v = Hd, (4.1)

This is the “expansion of the universe”. Eq. (4.1)

is Hubble’s law; H is Hubble’s constant. Its value

is 21 (km/sec)/million light years, within 5%.

An expanding spherical balloon, on which bits

of paper are glued, each representing a galaxy, pro-

vides a very instructive analogy. See Fig. 4.1. The

balloon’s two dimensional surface is the analog of

our universe’s three dimensional space. From the

viewpoint of every galaxy on the balloon, other

galaxies recede, and Eq. (4.1) holds. The galaxies are glued, not painted on,

because they do not expand with the universe.

We see the balloon expanding in an already existing three dimensional space,

but surface dwellers are not aware of this third spatial dimension. For both sur-

face dwellers in their universe and we in ours, galaxies separate not because they

67

4.1 Our Universe I

are moving apart in a static ﬁxed space, but because space itself is expanding,

carrying the galaxies with it.

We cannot verify Eq. 4.1 directly because neither d nor v can be measured

directly. But they can be measured indirectly, to moderate distances, as follows.

The distance d to objects at moderate distances can be determined from their

luminosity (energy received at Earth/unit time/unit area). If the object is near

enough so that relativistic eﬀects can be ignored, then its observed luminosity

at distance d is its absolute luminosity L (energy emitted/unit time) divided

by the area of a sphere of radius d:

=

L

4πd

2

. (4.2)

Thus the distance to the object can be determined from a measurement of its

luminosity (or the luminosity of a resolved star in it) if its absolute luminosity

is known. The absolute luminosity of various types of stars and galaxies has

been approximately determined, allowing an approximate determination of their

distances.

The velocity v of galaxies to moderate distances can be determined from their

redshift. Light from all galaxies, except a few nearby, exhibits a redshift z > 0.

By deﬁnition, z = f

e

/f

o

−1 (see Eq. (1.7)). If f

e

is known (for example if the

light is a known spectral line), then z is directly and precisely measurable. Eq.

(4.12) shows that the redshifts are not Doppler redshifts; they are ex:pansion

redshifts, the result of the expansion of space. However, to moderate distances

we can ignore relativistic eﬀects and determine v using the approximate Doppler

formula z = v from Exercise 1.6b.

Substituting z = v into Eq. (4.1) gives z = Hd, valid to moderate distances.

This approximation to Hubble’s law was established in 1929 by Edwin Hubble.

Since z increases with d, we can use z (which is directly and accurately measur-

able), as a proxy for d (which is not). Thus we say, e.g., that a galaxy is “at”

z = 2.

Reversing time, the universe contracts. From Eq. (4.1) galaxies approach us

with a velocity proportional to their distance. Thus they all arrive here at the

same time. Continuing into the past, the matter in the universe was extremely

compressed, and thus at an exceedingly high temperature, accompanied by ex-

ceedingly intense electromagnetic radiation. Matter and space were taking part

in an explosion, called the big bang. The recession of the galaxies that we see

today is the continuation of the explosion. The big bang is the origin of our

universe. It occurred 13.7 billion years ago. The idea of a hot early universe

preceding today’s expansion was ﬁrst proposed by Georges Lemaˆıtre in 1927.

There is no “site” of the big bang; it occurred everywhere. Our balloon again

provides an analogy. Imagine it expanding from a point, its big bang. At a later

time, there is no speciﬁc place on its surface where the big bang occurred.

Due to the ﬁnite speed of light, we see a galaxy not as it is today, but as it

was when the light we detect from it was emitted. When we look out in space,

we look back in time! This allows us to study the universe at earlier times. Fig.

68

4.1 Our Universe I

4.2, which is the graph of Eq. (4.21), illustrates this. Telescopes today see some

galaxies as they were when the universe was less than 1 billion years old.

In 1948 Ralph Alpher and Robert Her-

Fig. 4.2: The redshift z of a galaxy

vs. the age of the universe (in 10

9

years) when the light we see from it

was emitted.

man predicted that the intense electro-

magnetic radiation from the big bang,

much diluted and cooled by the expansion

of the universe, should still be with us to-

day. In 1965 Arno Penzias and Robert

Wilson, quite by accident and unaware

of the prediction, discovered microwave

radiation with the proper characteristics.

This cosmic background radiation (CBR)

provides strong evidence that a big bang

actually occurred. In Sec. 4.4 we shall see

that it contains a wealth of cosmological

information.

The CBR has a 2.7 K blackbody spec-

trum. The radiation started its journey

toward us ∼ 380, 000 years after the big

bang, when the universe had cooled to

about 3000 K, allowing nuclei and electrons to condense into neutral atoms,

which made the universe transparent to the radiation.

As a ﬁrst approximation the CBR is isotropic, i.e., spherically symmetric

about us. However, in 1977 astronomers measured a small dipole anisotropy: a

relative redshift in the CBR of z = −.0012 cos θ, where θ is the angle from the

constellation Leo.

Exercise 4.1. Explain this as due to the Earth moving toward Leo at 370

km/sec with respect to an isotropic CBR. In this sense we have discovered the

absolute motion of the Earth.

The relative redshift is modulated by the orbital motion of the Earth around

the Sun. After correcting for the dipole anisotropy, the CBR is isotropic to 1

part in 10,000.

Further evidence for a big bang comes from the theory of big bang nucleosyn-

thesis. The theory predicts that from a few seconds to a few minutes after the

big bang, nuclei formed, almost all hydrogen (75% by mass) and helium-4 (25%),

with traces of some other light elements. Observations conﬁrm the prediction.

This is a spectacular success of the big bang theory. Other elements constitute

only a small fraction of the mass of the universe. They are mostly formed in

stars and are strewn into interstellar space by supernovae and by other means,

to be incorporated into new stars, their planets, and any life which may arise

on the planets.

The big bang is part of the most remarkable generalization of science: The

universe had an origin and is evolving at many interrelated levels. The lev-

els include galactic evolution, stellar evolution, planetary evolution, biological

evolution, and cultural evolution.

69

4.2 Robertson-Walker Spacetimes

4.2 Robertson-Walker Spacetimes

The CBR is highly isotropic. Galaxies are distributed isotropically. Redshift

surveys of hundreds of thousands of galaxies show that the universe is homoge-

neous in space. Our general relativistic model ignores the clumping of matter

into galaxies by assuming an exactly isotropic and homogeneous continuous dis-

tribution of inertial matter in the universe. This should not aﬀect the large

scale structure of the universe, which is what we are interested in.

We ﬁrst set up a coordinate system (t, r, φ, θ) for the universe. As the uni-

verse expands, the density ρ of matter decreases. Place a clock at rest in every

galaxy and set it to some agreed time when some agreed value of ρ is observed

at the galaxy. The t coordinate is the (proper) time measured by these clocks.

Choose any galaxy as the spatial origin. Isotropy demands that the φ and θ

coordinates of other galaxies do not change in time. Choose any time t

0

. Deﬁne

the r coordinate at t

0

to be any quantity that increases with increasing distance

from the origin. Galaxies in diﬀerent directions but at the same distance are

assigned the same r. Now deﬁne the r coordinate of galaxies at times other

than t

0

by requiring that r, like φ and θ, does not change. Isotropy ensures that

all galaxies at a given r at a given t are at the same distance from the origin.

The coordinates (r, φ, θ) are called comoving. Distances between galaxies

change in time not because their (r, φ, θ) coordinates change, but because their

t coordinate changes, which changes the metric.

Our balloon provides an analogy. If the radius of the balloon at time t is

S(t), then according to Eq. (2.11) the balloon’s metric at time t is

ds

2

= S

2

(t) d Ω

2

. (4.3)

If S(t) increases, then the balloon expands. Distances between the glued on

galaxies increase, but their comoving (φ, θ) coordinates do not change.

Exercise 4.2. Derive the analog of Eq. (4.1) for the balloon. Let d be the

distance between two galaxies on the balloon at time t, and let v be the speed

with which they are separating. Show that v = Hd, where H = S

(t)/S(t).

Hint: See the derivation of Eq. (4.10).

Exercise 4.3. Substitute sinφ = r/(1 +r

2

/4) in Eq. (4.3). Show that the

metric with respect to the comoving coordinates (r, θ) is

ds

2

= S

2

(t)

dr

2

+r

2

dθ

2

(1 +r

2

/4)

2

. (4.4)

Again, distances between the glued on galaxies increase with S(t) , but their

comoving (r, θ) coordinates do not change.

70

4.2 Robertson-Walker Spacetimes

The metric of an isotropic universe, after suitably choosing the comoving r

coordinate, is:

Robertson-Walker Metric

ds

2

= dt

2

−S

2

(t)

dr

2

+r

2

d Ω

2

(1 +k r

2

/4)

2

(k = 0, ±1)

≡ dt

2

−S

2

(t) dσ

2

. (4.5)

The elementary but somewhat involved derivation is given in Appendix 14. The

ﬁeld equation is not used. We will use it in Sec. 4.4 to determine S(t). The

constant k is the curvature of the spatial metric dσ: k = −1, 0, 1 for negatively

curved, ﬂat, and positively curved respectively.

The metric was discovered independently by H. P. Robertson and E. W.

Walker in the mid 1930’s. It should be compared with the metric in Eq. (4.4).

Emit a pulse of light from a galaxy to a neighboring galaxy. Since ds = 0 for

light, Eq. (4.5) gives dt = S(t) dσ. By deﬁnition, or by the Robertson-Walker

metric Eq. (4.5), t is measured by clocks at rest in galaxies. Thus dt is the

elapsed time in a local inertial frame in which the emitting galaxy is at rest.

Since c = 1 in local inertial frames, the physical distance between the galaxies

is

S(t) dσ. (4.6)

This is the distance that would be measured by a rigid rod. Since dσ does not

change in time, the distance is proportional to S(t), which is thus a scale factor

for the universe.

The metric of points at coordinate radius r in the Robertson-Walker metric

Eq. (4.5) is that of a sphere of radius S(t) r/(1 + k r

2

/4). From Eq. (4.6) this

metric measures physical distances. Thus the sphere has surface area

A(r) =

4πr

2

S

2

(t)

(1 +k r

2

/4)

2

(4.7)

and volume

V (r) = S(t)

r

0

A(ρ)

1 +kρ

2

/4

dρ . (4.8)

If k = −1 (negative spatial curvature), then r is restricted to 0 ≤ r < 2. The

substitution ρ = 2 tanh ξ shows that V (2) = ∞, i.e., the universe has inﬁnite

volume. If k = 0 (zero spatial curvature), 0 ≤ r < ∞ and the universe has

inﬁnite volume. If k = 1 (positive spatial curvature), then A(r) increases as r

increases from 0 to 2, but then decreases to zero as r → ∞. The substitution

ρ = 2 tan ξ shows that V (∞) = 2π

2

S

3

(t); the universe has ﬁnite volume! The

surface of our balloon provides an analogy. It is ﬁnite and has positive curvature.

Circumferences of circles of increasing radius centered at the North pole increase

until the equator is reached and then decrease to zero.

71

4.3 The Expansion Redshift

4.3 The Expansion Redshift

At time t the distance from Earth at r = 0 to a galaxy at r = r

e

is, by Eq.

(4.6),

d = S(t)

re

0

dσ. (4.9)

This is the distance that would be obtained by adding the lengths of small rigid

rods laid end to end between Earth and the galaxy at time t.

Diﬀerentiate to give the velocity of the galaxy with respect to the Earth:

v = S

(t)

re

0

dσ.

Divide to give Hubble’s law v = Hd, Eq. (4.1), where

H =

S

(t)

S(t)

. (4.10)

From v = Hd we see that if d = 1/H ≈ 13.8 10

9

light years, then v = 1 ,

the speed of light. Galaxies beyond this distance, at z ≈ 1.5 , are receding

from us faster than the speed of light. What about the rule “nothing can move

faster than light”? In special relativity the rule applies to objects moving in

an inertial frame, and in general relativity to objects moving in a local inertial

frame. There is no local inertial frame containing both the Earth and the galaxy.

We now obtain a relationship between the expansion redshift of light and

the size of the universe when the light was emitted. Suppose that light signals

are emitted toward us at events (t

e

, r

e

) and (t

e

+∆t

e

, r

e

) and received by us at

events (t

o

, 0) and (t

o

+ ∆t

o

, 0). Since ds = 0 for light, Eq. (4.5) gives

dt

S(t)

=

dr

1 +k r

2

/4

. (4.11)

Integrate this over both light worldlines to give

to

te

dt

S(t)

=

re

0

dr

1 +k r

2

/4

=

to+∆to

te+∆te

dt

S(t)

.

Subtract

to

te+∆te

from both ends to give

te+∆te

te

dt

S(t)

=

to+∆to

to

dt

S(t)

.

72

4.3 The Expansion Redshift

If ∆t

e

and ∆t

o

are small, this becomes

∆t

e

S(t

e

)

=

∆t

o

S(t

o

)

.

Use Eq. (1.6) to obtain the desired relation:

z + 1 =

S(t

o

)

S(t

e

)

. (4.12)

This wonderfully simple formula directly relates the expansion redshift to the

universe’s expansion. Since S = 0 at the big bang, the big bang is at z = ∞.

A galaxy at redshift z ≈ 10 is the most distant known. The light we see

from the galaxy was emitted ∼ 470 million years after the big bang. The faint

galaxy is magniﬁed 25-100 times by a gravitational lensing galaxy at z = .25.

According to Eq. (4.12) the galaxy is 11 times farther away from us today

than when it emitted the light we detect from it! If we could observe a physical

process on the galaxy from Earth, then by Eq. (1.6) it would appear to proceed

at 1/11 of the rate of the same process on Earth.

Several distant supernovae have exhibited this eﬀect. For example, a super-

nova at z = .5 appeared to last about 1 + .5 = 1.5 times longer than similar

supernovae nearby.

Suppose that an object at distance z emits blackbody radiation at temper-

ature T(z). According to Planck’s law, the object radiates at frequency f

e

with

intensity proportional to

f

3

e

e

hfe/kT(z)

−1

,

where h is Planck’s constant and k is Boltzmann’s constant. Radiation emitted

at f

e

is received by us redshifted to f

o

, where from Eq. (1.7), f

e

= (z + 1)f

o

.

Thus the intensity of the received radiation is proportional to

f

3

o

e

hfo/kT(0)

−1

,

where T(0) is deﬁned by

z + 1 =

T(z)

T(0)

. (4.13)

The observed spectrum is also blackbody, at temperature T(0).

As stated in Sec. 4.1, the CBR was emitted at ≈ 3000 K and is observed

today at 2.7 K. Thus from Eq. (4.13), its redshift z ≈ 3000K/2.7K ≈ 1100.

Analysis of spectral lines in a quasar at z = 3 shows that they were excited

by an ∼11 K CBR. Eq. (4.13) predicts this: T(3) = (3 + 1)T(0) ≈ 11K. From

Eq. (4.12), the universe was then about 1/(1 + 3) = 25% of its present size.

73

4.4 Our Universe II

4.4 Our Universe II

The Robertson-Walker metric Eq. (4.5) is a consequence of isotropy alone; the

ﬁeld equation was not used in its derivation. There are two unknowns in the

metric: the curvature k and the expansion factor S(t). In this section we ﬁrst

review the compelling evidence that in our universe k = 0. This spatial ﬂatness,

together with observational data, will force us to modify the ﬁeld equation. We

then solve the modiﬁed ﬁeld equation to determine S(t).

In 1981 Alan Guth proposed that there was a period of exponential expan-

sion, called inﬂation, in the very early universe. The inﬂation started perhaps

10

−34

sec after the big bang, lasted perhaps 10

−35

sec, and expanded the uni-

verse by a factor of 10

22

! As fantastic as inﬂation seems, it explained several

puzzling cosmological observations and steadily gained support over the years.

By 1990, most cosmologists believed that inﬂation occurred. Since then more

evidence for inﬂation has accumulated. Inﬂation predicts a ﬂat universe.

There is more direct evidence for a ﬂat universe. When the CBR was emit-

ted, the density of matter was not quite uniform. Regions of higher density

led to the galaxy clusters we wee today. Those regions were hotter, leading

to tiny temperature anisotropies (∼ 60 mK, a few parts in 10

5

) in the CBR.

Calculations show that in a ﬂat universe, the average temperature diﬀerence as

a function of angular separation peaks at .8 arcdegree. The angle is diﬀerent for

a curved space. Observations in 2000 found the peak at .8 arcdegree, providing

compelling evidence for a ﬂat universe.

Fig. 4.3: Angular size

on a sphere.

To understand how curvature can aﬀect angular size,

consider the everyday experience that the angular size of

an object decreases with distance. The rate of decrease

is diﬀerent for diﬀerently curved spaces. Imagine surface

dwellers whose universe is the surface of the sphere in

Fig. 4.3 and suppose that light travels along the great

circles of the sphere. According to Exercise 2.9 these

circles are geodesics. For an observer at the north pole,

the angular size α of an object of length D decreases

with distance, but at a slower rate than on a ﬂat surface.

And when the object reaches the equator, its angular

size begins to increase. The object has the same angular

size at the two positions in the ﬁgure. On a pseudosphere the angular size

decreases at a faster rate than on a ﬂat surface.

Despite the strong evidence for a ﬂat universe, we shall see that an applica-

tion of the ﬁeld equation Eq. (2.27) to the Robertson-Walker metric Eq. (4.5)

with k = 0 gives results in conﬂict with observation. The simplest solution to

this problem is to add a term Λg to the ﬁeld equation:

G+ Λg = −8πκT. (4.14)

This is the most natural change possible to the ﬁeld equation. In fact, we have

seen the term Λg before, in the ﬁeld equation Eq. (2.34). We eliminated it

74

4.4 Our Universe II

there by assuming that spacetime is ﬂat in the absence of matter. Thus the

term represents a gravitational eﬀect of empty space. The constant Λ is called

the cosmological constant. If it is small enough, then its eﬀects in the solar

system are negligible, but it can dramatically aﬀect cosmological models.

To apply the new ﬁeld equation, we need the energy-momentum tensor T.

In Sec. 4.2 we distributed the matter in the universe into a uniform distribution

of density ρ. We assume that the matter is dust. (This is certainly reasonable

today and certainly unreasonable in the very early universe.) Thus T is of the

form Eq. (2.26). For the comoving matter, dr = dθ = dφ = 0, and from Eq.

(4.5), dt/ds = 1. Thus T

tt

= ρ is the only nonzero T

jk

.

The ﬁeld equation Eq. (4.14) for the Robertson-Walker metric Eq. (4.5) with

k = 0 and the T just obtained reduces to two ordinary diﬀerential equations:

−3

S

S

2

+ Λ = −8πκρ , (tt component) (4.15)

S

2

+ 2 SS

−ΛS

2

= 0 . (rr, φφ, θθ components) (4.16)

From Eqs. (4.15) and (4.10),

ρ +

Λ

8πκ

= ρ

c

, (4.17)

where ρ

c

≡ 3H

2

/8πκ is called the critical density. Today ρ

c

≈ 10

−29

g/cm

3

–

the equivalent of a few hydrogen atoms per cubic meter.

Precise measurements of the CBR by the Wilkinson Microwave Anisotropy

Probe (WMAP) were announced in 2003. Anisotropies in the CBR of a few

parts in 10

5

were measured. The WMAP results determined the value of sev-

eral cosmological parameters reported in this section. The clustering of galaxies

is sensitive to the value of some of these parameters. Measurements of the posi-

tions and redshifts of 200,000 galaxies by the Sloan Digital Sky Survey (SDSS)

were also announced in 2003. The SDSS results give a second, independent,

determination of the parameters, which agree with those of WMAP.

The measurements show that today the density of matter is ρ = .3ρ

c

. This

conﬁrmed earlier estimates based on gravitational lens statistics. Some of this

matter is in ordinary known forms (neutrons, protons, electrons, neutrinos,

photons, . . . ), but most is not.

The measurements show that the known forms of matter constitute .05ρ

c

.

This conﬁrmed two earlier determinations. One used spectra of distant galac-

tic nuclei to “directly” measure the density of known forms of matter in the

early universe. The other inferred the density from the abundance of deuterium

created during big bang nucleosynthesis. The deuterium abundance was mea-

sured in distant pristine gas clouds backlit by even more distant quasars. It is

gratifying to have excellent agreement between the CBR and deuterium mea-

surements, as the deuterium was created a few minutes after the big bang the

CBR was emitted hundreds of thousands of years later. Today as much as 2/3

of this matter is in vast superhot intergalactic gas clouds.

75

4.4 Our Universe II

The measurements show that unknown forms of matter constitute .25ρ

c

.

This matter does not interact electromagnetically. It is therefore called dark

matter. Gravitationally, dark matter behaves just as ordinary matter. The

existence and approximate abundance of dark matter was ﬁrst inferred from its

gravitational eﬀects. For example, stars orbiting in the outer part of a galaxy

move too rapidly for the density of ordinary matter to hold them in orbit.

Gravitational lensing is used to map the distribution of dark matter. In one

system, a quasar is lensed by a foreground cluster of galaxies. The separation of

the quasar images is so large that dark matter in the cluster must be responsible

for the lensing.

Dark matter explains the growth of structure in our universe; galaxies could

not have formed without it. Galaxies arose from small random density ﬂuc-

tuations in the very early very hot universe. Density ﬂuctuations of ordinary

matter were smoothed out by interactions with electromagnetic radiation. But

dark matter does not interact electromagnetically, and so density ﬂuctuations

within it could grow.

From Eq. (4.17), Λ/8πκ = ρ

c

− .3ρ

c

= .7ρ

c

. The quantity Λ/8πκ is the

density of another mysterious component of our universe, called dark energy.

Dark energy is intrinsic to space; it literally “comes with the territory”. This

“it is what is left over in Eq. (4.17)” argument for its existence is perhaps

unsatisfying. We shall see soon that there is observational evidence for dark

energy.

Exercise 4.4. Show that Λ ≈ 10

−35

/sec

2

.

Fig. 4.4: S (in units of S(t0)) as a

function of t (in 10

9

years).

We now solve the ﬁeld equations (4.15)

and (4.16) for S. A quick calculation using

Eq. (4.16) shows that

SS

2

−ΛS

3

/3

=

0. Thus from Eq. (4.15),

ρ S

3

= constant. (4.18)

According to Eq. (4.18), ρ is inversely pro-

portional to the cube of the scale factor S.

This is what we would expect. In contrast,

the density of dark energy, Λ/8πκ, is con-

stant in time (and space). Thus the ratio of

the dark energy density to the matter den-

sity increases with time. At early times mat-

ter dominates and at later times dark energy

dominates.

Now multiply Eq. (4.15) by −S

3

and use Eq. (4.18):

3SS

2

−ΛS

3

= 8πκρ

0

S

3

(t

0

) . (4.19)

Integrate this to solve the ﬁeld equations:

S

3

(t) =

8πκρ

0

Λ

sinh

2

√

3Λ

2

t

S

3

(t

0

). (4.20)

76

4.4 Our Universe II

Since the big bang is the origin of the universe, it is convenient to set t = 0

when S = 0. This convention eliminates a constant of integration in Eq. (4.20).

The graph of S(t) is given in Fig. 4.4.

Exercise 4.5. Show that Eq. (4.20) implies that the age of the universe is

t

0

≈ 13.7 billion years.

The graph of S(t) begins concave down, i.e., the expansion slows down. This

is due to the gravitational attraction of matter, which dominates at early times.

Then the graph becomes concave up, i.e., the expansion speeds up! This is

due to the dark energy Λ, which acts as a repulsive force, becoming dominant.

Gravity is repulsive on cosmological scales today!

Exercise 4.6. Show that the transition from concave down to concave up

occurred ∼5 billion years ago when the universe was ∼2/3 of its present size.

Exercise 4.7. Show that Eq. (4.20) implies that in the future the universe

will undergo exponential expansion.

Using Eqs. (4.12) and (4.20) we can relate the redshift z of a galaxy to the

time t it emitted the light we see from it:

(z + 1)

−3

=

8πκρ

0

Λ

sinh

2

√

3Λ

2

t

. (4.21)

The graph of z vs. t was given in Fig. 4.2. The transition from the expansion

slowing down to speeding up occurred at z ≈ .5.

We are now in a position to discuss the

Fig. 4.5: Luminosity-redshift rela-

tion. Upper: ρ/ρc = 1; middle:

ρ/ρc = .3; lower: ρ/ρc = 0.

observational evidence for dark energy. The

best evidence comes from observations of

distant supernovae. All type Ia supernovae

have about the same peak intrinsic lumi-

nosity L. The luminosity-redshift relation,

derived in Appendix 15, expresses the ob-

served luminosity of an object as a func-

tion of its intrinsic luminosity and redshift.

The function also depends on the current

value of ρ/ρ

c

. Fig. 4.5 is the graph of for a

ﬁxed L and three values of ρ/ρ

c

. The value

ρ/ρ

c

= 1 corresponds to Λ = 0 . The points

(z, ) for supernovae in the range z ≈ .5 to

z ≈ 1.5 lie closest to the ρ/ρ

c

= .3 curve,

and deﬁnitely not the Λ = 0 curve. This is

strong evidence that Λ = 0 . The eﬀect of Λ is dramatic: the luminosities of

z ≈ .5 supernovae are ∼30% less than they would be if Λ were 0.

77

4.4 Our Universe II

Further evidence for dark energy comes from observations of the CBR in

directions near galaxy clusters. As CBR enters the gravitational ﬁeld of a cluster

it gains energy, just as a material object would. When leaving a static ﬁeld the

radiation loses exactly the energy gained. But in a universe accelerating due to

dark energy, the ﬁeld of the cluster weakens during the light’s transit through

it and the loss is not quite as large as the gain. From Planck’s law E = hf and

the redshift formula Eq. (1.7), the light suﬀers a relative redshift z < 0 . This

integrated Sachs-Wolfe eﬀect was detected in 2003.

Summary. The universe had an origin 13.7 billion years ago. It is ﬂat,

expanding, and the expansion is speeding up. It consists of 5% known matter,

25% dark matter, and 70% dark energy. Each of these numbers is supported

by independent lines of evidence.

The construction of a cosmological model which ﬁts the observations so

well is a magniﬁcent achievement. But it crystallized only at the turn of the

millennium. And we are left with two mysteries: the nature of dark matter and

dark energy. They constitute 95% of the “stuﬀ” of the universe! We are in “a

golden age of cosmology” in which new cosmological data are pouring in. We

must be prepared for changes, even radical changes, in our model.

The Horizon. Consider light emitted at the big bang at event (t = 0, r =

r

h

) and received by us today at event (t = t

0

, r = 0). Our horizon consists of

points at r = r

h

. Galaxies beyond the horizon cannot be seen by us because

their light has not had time to reach us since the big bang! Conversely, galaxies

beyond the horizon cannot see us.

From Eq. (4.9), with dσ = dr for k = 0, the distance to r

h

today is d

h

=

S(t

0

) r

h

. Integrate Eq. (4.11) along the light worldline to obtain

d

h

= S(t

0

)

t0

0

dt

S(t)

. (4.22)

Substituting for S(t) from Eq. (4.20) gives d

h

≈ 47 billion light years.

As t

0

increases, so does d

h

, and so we can see, and can be seen by, more

galaxies with passing time. However, the right side of Eq. (4.22) converges as

t

0

→∞and so there are galaxies which will remain beyond the horizon forever.

Today d

h

is about 3/4 of its ultimate value.

Loose End. We close this section by tying up a loose end from Chapter 2:

We show that A = 1 in Eq. (2.34). Set Λ = 0 in the ﬁeld equations, Eqs. (4.15)

and (4.16), solve Eq. (4.16) for S

, and use Eq. (4.15):

S

= −

4πκρS

3

. (4.23)

Consider a ball of inertial mather of density ρ centered at the origin. Using

Eq. (4.6), the ball has radius S(t)

dσ = S(t) σ (σ is constant). According

to the discussion following Exercise 3.3 (which is based on the Schwartzschild

metric and therefore only the vacuum ﬁeld equation), we may, without aﬀecting

the motion of a particle on the surface of the ball, (1) ignore the matter outside

78

4.4 Our Universe II

the ball, and (2) shrink the matter inside the ball in a spherically symmetric

way to a small ball at the center. The same is true of Newtonian gravitation.

Thus if S(t) σ is small, then we may apply the Newtonian equation Eq. (2.1) to

the motion of the particle:

S

σ = −κ

4

3

π (Sσ)

3

ρ

(Sσ)

2

,

which is the same as Eq. (4.23). Einstein’s and Newton’s theories give the same

result in this situation.

But if A = 1 in Eq. (2.34), then the left side of Eq. (4.23) would be

multiplied by A. Thus the coincidence of the two theories in this situation

requires A = 1.

79

4.5 General Relativity Today

4.5 General Relativity Today

General relativity extended the revolution in our ideas about space and time

begun by special relativity. Special relativity showed that space and time are

not absolute and independent, but related parts of a whole: spacetime. Gen-

eral relativity showed that a spacetime containing matter is not ﬂat and static,

but curved and dynamic, not only aﬀecting matter (according to the geodesic

postulate), but also aﬀected by matter (according to the ﬁeld equation). In con-

sequence, Euclidean geometry, thought inviolate for over twenty centuries, does

not apply (exactly) in a gravitational ﬁeld. The Robertson-Walker spacetimes

illustrate these points particularly clearly.

Despite its fundamental nature, general relativity was for half a century

after its discovery in 1915 out of the mainstream of physical theory because its

results were needed only for the explanation of the few minute eﬀects in the

solar system described in Sec. 3.3. Newton’s theory of gravity accounted for all

other gravitational phenomena. But since 1960 astronomers have made several

spectacular discoveries: quasars, the CBR, neutron stars, the binary pulsar,

gravitational lenses, the emission of gravitational radiation, black holes, detailed

structure in the CBR, and that the expansion of the universe is speeding up.

These discoveries have revealed a wonderfully varied, complex, and interesting

universe. They have enormously widened our “cosmic consciousness”. And we

need general relativity to help us understand them.

General relativity has passed every test to which it has been put. General

relativity reduces to special relativity, an abundantly veriﬁed theory, in the

absence of signiﬁcant gravitational ﬁelds. General relativity reduces to Newton’s

theory of gravity, an exceedingly accurate theory, in weak gravity in which the

velocity of matter is small. Sec. 2.2 cited experimental evidence and logical

coherence as reasons for accepting the postulates of the theory. The theory

explains the minute details of the motion of matter and light in the solar system;

the tests of the vacuum ﬁeld equation described in Sec. 3.3 – the perihelion

advance of Mercury, light deﬂection, and radar echo delay – all conﬁrm the

predictions of general relativity with an impressive accuracy. The lunar laser

experiment conﬁrms tiny general relativistic eﬀects on the motion of the Earth

and moon. Gravitomagnetism has probably been detected. The perihelion

advance of the binary pulsar and the apparent emission of gravitational radiation

from its system provide quantitative conﬁrmation of the full ﬁeld equation.

Finally, the application of general relativity on the grandest possible scale, to

the universe as a whole, appears to be successful.

As satisfying as all this is, we would like more evidence for such a fundamen-

tal theory. This would give us more conﬁdence in applying it. The diﬃculty in

ﬁnding tests for general relativity is that the gravitational interaction is weak

(about 10

−40

as strong as the electromagnetic interaction), and so astronom-

ical sized masses, over which an experimenter has no control, must be used.

Also, Newton’s theory is already very accurate in most circumstances. Rapidly

advancing technology has made, and will continue to make, more tests possible.

80

4.5 General Relativity Today

Foremost among these are attempts to detect gravitational radiation. The

hope is that this radiation will someday complement electromagnetic radiation

in providing information about the universe. Sec. 3.5 described observations

of the binary pulsar which show that gravitational radiation exists. Several

ground based detectors are in operation, and several are planned in space, to

try to detect it directly. We still await the ﬁrst such detection.

The most unsatisfactory feature of general relativity is its conception of mat-

ter. The density ρ in the energy-momentum tensor T represents a continuous

distribution of matter, entirely ignoring its atomic and subatomic structure.

Quantum theory describes this structure. The standard model is a quantum

theory of subatomic particles and forces. The theory, forged in the 1960’s and

1970’s, describes all known forms of matter, and the forces between them, except

gravity. The fundamental particles of the theory are quarks, leptons, and force

carriers. There are several kinds of each. Protons and neutrons consist of three

quarks. Electrons and neutrinos are leptons. Photons carry the electromagnetic

force.

The standard model has been spectacularly successful, correctly predict-

ing the results of all particle accelerator experiments (although many expect

discrepancies to appear soon in new experiments). The theory of big bang nu-

cleosynthesis uses the standard model. However, in cosmology the standard

model leaves many matters unexplained. For example, inﬂation, dark matter,

and dark energy are all beyond the scope of the theory.

Quantum theory and general relativity are the two fundamental theories

of contemporary physics. But they are separate theories; they have not been

uniﬁed. Thus they must be combined on an ad hoc basis when both gravity

and quantum eﬀects are important. For example, in 1974 Steven Hawking

used quantum theoretical arguments to show that a black hole emits subatomic

particles and light. A black hole is not black! The eﬀect is negligible for a black

hole with the mass of the Sun, but it would be important for very small black

holes.

The uniﬁcation of quantum theory and general relativity is the most impor-

tant goal of theoretical physics today. Despite decades of intense eﬀort, the goal

has not been reached. String theory and loop quantum gravity are candidates

for a theory of quantum gravity, but they are very much works in progress.

Despite the limited experimental evidence for general relativity and the gulf

between it and quantum theory, it is a necessary, formidable, and important tool

in our quest for a deeper understanding of the universe in which we live. This

and the striking beauty and simplicity – both conceptually and mathematically

– of the uniﬁcation of space, time, and gravity in the theory make general

relativity one of the ﬁnest creations of the human mind.

81

Appendices

A.1 Physical Constants. When appropriate, the ﬁrst number given uses

(light) seconds as the unit of distance.

c = 1 = 3.00 10

10

cm/sec = 186,000 mi/sec (Light speed)

κ = 2.47 10

−39

sec/g = 6.67 10

−8

cm

3

/(g sec

2

)

(Newtonian gravitational constant)

g = 3.27 10

−8

/sec = 981cm/sec

2

= 32.2 ft/sec

2

(Acceleration Earth gravity)

Mass Sun = 1.99 10

33

g

Radius Sun = 2.32 sec = 6.96 10

10

cm = 432,000 mi

Mean radius Mercury orbit = 1.9 10

2

sec = 5.8 10

12

cm = 36,000,000 mi

Mercury perihelion distance = 1.53 10

2

sec = 4.59 10

12

cm

Eccentricity Mercury orbit = .206

Period Mercury = 7.60 10

6

sec = 88.0 day

Mean radius Earth orbit = 5 10

2

sec = 1.5 10

13

cm = 93,000,000 mi

Mass Earth = 5.97 10

27

g

Radius Earth = 2.13 10

−2

sec = 6.38 10

8

cm = 4000 mi

Angular momentum Earth = 10

20

g sec = 10

41

g cm

2

/sec

H

0

= 2.3 10

−18

/sec = (22 km/sec)/10

6

light year (Hubble constant)

A.2 Approximations. The approximations are valid for small x.

(1 +x)

n

≈ 1 +nx

e.g., (1 +x)

1

2

≈ 1 +x/2 and (1 +x)

−1

≈ 1 −x

sin(x) ≈ x −x

3

/6

cos(x) ≈ 1

82

A.3 The Macek-Davis Experiment

A.3 The Macek-Davis Experiment. In this experiment, light was

reﬂected in both directions around a ring laser, setting up a standing wave. A

fraction of the light in both directions was extracted by means of a half silvered

mirror and their frequencies compared. See Fig. A.6.

Let t

1

and t

2

be the times it takes light

Fig. A.6: Comparing the round trip

speed of light in opposite directions.

to go around in the two directions and f

1

and f

2

be the corresponding frequencies.

Then t

1

f

1

= t

2

f

2

,, as both are equal to

the number of wavelengths in the stand-

ing wave. Thus any fractional diﬀerence

between t

1

and t

2

would be accompanied

by an equal fractional diﬀerence between f

1

and f

2

. The frequencies diﬀered by no more

than one part in 10

12

.

(If the apparatus is rotating, say, clock-

wise, then the clockwise beam will have a

longer path, having to “catch up” to the

moving mirrors, and the frequencies of the

beams will diﬀer. The experiment was performed to detect this eﬀect.)

83

A.4 Moving Rods

A.4 Moving Rods. If an inertial rod is pointing perpendicular to its

direction of motion in an inertial frame, then its length, as measured in the

inertial frame, is unchanged. If the rod is pointing parallel to its direction

of motion in the inertial frame, then its length, as measured in the inertial

frame, contracts. Why this diﬀerence? An answer can be given in terms of

Einstein’s principle of relativity: identical experiments performed in diﬀerent

inertial frames give identical results. We show that the principle demands that

the length be unchanged in the perpendicular case, but not in the parallel case.

Consider identical rods R and R

**, both inertial and in motion with respect
**

to each other. According to the inertial frame postulate, R and R

are at rest

in inertial frames, say I and I

.

Fig. A.7: Rod

perpendicular

to its direc-

tion of motion

in I.

Perpendicular case. First, situate the rods as in Fig. A.7.

Suppose also that the ends A and A

**coincide when the rods
**

cross. Then we can compare the lengths of the rods directly,

independently of any inertial frame, by observing the relative

positions of the ends B and B

**when the rods cross. According
**

to the principle of relativity, R and R

**have the same length. For
**

if, say, B

**passed below B, then this would violate the principle:
**

a rod moving in I is shorter, while an identical rod moving in I

is longer.

Parallel case. Now situate the rods as in Fig. A.8. Wait

until A and A

**coincide. Consider the relative positions of B
**

and B

**at the same time. But “at the same time” is diﬀerent
**

in I and I

! What happens is this: At the same time in I, B

will not have reached B, so R

**is shorter than R in I. And at
**

the same time in I

, B will have passed B

, so R is shorter than

R

in I

**. This is length contraction. There is no violation of the
**

principle of relativity here: in each frame the moving rod is shorter.

Fig. A.8: Rod parallel to its direction of

motion in I.

There is no contradiction here: The

lengths are compared diﬀerently in the

two inertial frames, each frame using the

synchronized clocks at rest in that frame.

Thus there is no reason for the two com-

parisons to agree. This is diﬀerent from

Fig. A.7, where the lengths are compared directly. Then there can be no dis-

agreement over the relative lengths of the rods.

We calculate the length contraction of R

in I. Let the length of R

in I and

I

be D and D

, respectively. Let v be the speed of R

in I. Emit a pulse of

light from A

toward B

. In I, B

is at distance D from A

**when the light is
**

emitted.

Since c = 1 in I, the relative speed of the light and B

in I is 1 − v. Thus

in I the light will take time D/(1 − v) to reach A

**. Similarly, if the pulse is
**

reﬂected at B

back to A

, it will take time D/(1 + v) to reach A

. Thus the

84

A.4 Moving Rods

round trip time in I is

∆x

0

=

D

1 −v

+

D

1 +v

=

2D

1 −v

2

.

According to Eq. (1.12), a clock at A

**will measure a total time
**

∆s =

1 −v

2

1

2

∆x

0

for the pulse to leave and return.

Since c = 1 in I

and since the light travels a distance 2D

in I

, ∆s is also

given by

∆s = 2D

.

Combine the last three equations to give D = (1 −v

2

)

1

2

D

; the length of R

,

as measured in I, is contracted by a factor (1 −v

2

)

1

2

of its rest length.

85

A.5 The Michelson-Morley Experiment

A.5 The Michelson-Morley Experiment. Michelson and Morley used

a Michelson interferometer that splits a beam of monochromatic light in per-

pendicular directions by means of a half-silvered mirror. The beams reﬂect oﬀ

mirrors and return to the half-silvered mirror, where they reunite and proceed

to an observer. See Fig. A.9.

Let 2T be the time for the light to traverse

Fig. A.9: A Michelson interfer-

ometer.

an arm of the interferometer and return, D the

length of the arm, and c the two way speed of

light in the arm. Then by the deﬁnition of two

way speed, 2T = 2D/c. The diﬀerence in the

times for the two arms is 2D

1

/c

1

−2D

2

/c

2

. If

f is the frequency of the light, then the light

in the two arms will reunite

N = f

2D

1

c

1

−

2D

2

c

2

(A.1)

cycles out of phase.

If N is a whole number, the uniting beams

will constructively interfere and the observer

will see light; if N is a whole number +

1

2

, the uniting beams will destructively

interfere and the observer will not see light.

Rotating the interferometer 90

◦

switches c

1

and c

2

, giving a new phase dif-

ference

N

= f

2D

1

c

2

−

2D

2

c

1

.

As the interferometer rotates, light and dark will alternate N − N

times. Set

c

2

−c

1

= ∆c and use c

1

≈ c

2

= c (say) to give

N −N

= 2f (D

1

+D

2

)

1

c

1

−

1

c

2

≈ 2f

D

1

+D

2

c

∆c

c

. (A.2)

In Joos’ experiment, f = 6 10

14

/sec, D

1

= D

2

= 2100 cm, and [N −N

[ <

10

−3

. From Eq. (A.2),

∆c

c

< 6 10

−12

.

86

A.6 The Brillit-Hall Experiment

A.6 The Brillit-Hall Experiment. In this experiment, the output of

a laser was reﬂected between two mirrors, setting up a standing wave. The

frequency of the laser was servostabilized to maintain the standing wave. This

part of the experiment was placed on a granite slab that was rotated. Fig. A.10

shows a schematic diagram of the experiment.

Let 2T be the time for light to

Fig. A.10: The Brillit-Hall experiment.

travel from one of the mirrors to the

other and back, f be the frequency

of the light, and N be the num-

ber of wavelengths in the standing

wave. The whole number N is held

constant by the servo. Then

Tf = N. (A.3)

But 2T = 2D/c, where D is the

distance between the mirrors and c

is the two way speed of light be-

tween the mirrors. Substitute this

into Eq. (A.3) to give Df = cN.

Thus any fractional change in c as the slab rotates would be accompanied by

an equal fractional change in f. The frequency was monitored by diverting a

portion of the light oﬀ the slab and comparing it with the output of a reference

laser which did not rotate. The fractional change in f was no more than four

parts in 10

15

.

87

A.7 The Kennedy-Thorndike Experiment

A.7 The Kennedy-Thorndike Experiment. Kennedy and Thorndike

used a Michelson interferometer with arms of unequal length. Instead of rotating

the interferometer to observe changes in N in they observed the interference of

the uniting beams over the course of several months. Any change dc in the

two way speed of light (due, presumably, to a change in the Earth’s position or

speed) would result in a change dN in N: if we set, according to the result of

the Michelson-Morley experiment, c

1

= c

2

= c in Eq. (A.1), we obtain

dN = 2f

D

1

−D

2

c

dc

c

In the experiment, f = 610

14

/sec, D

1

−D

2

= 16 cm, and [dN[ < 310

−3

.

Thus [dc/c[ < 6 10

−9

.

88

A.8 Newtonian Orbits

A.8 Newtonian Orbits. Let the Sun be at the origin and r(t) be the path

of the planet. At every point r deﬁne unit vectors

Fig. A.11: The polar unit

vectors.

u

r

= cos θ i + sin θ j ,

u

θ

= −sin θ i + cos θ j .

See Fig. 8. Diﬀerentiate r = ru

r

to give the ve-

locity

v =

dr

dt

=

dr

dt

u

r

+r

du

r

dθ

dθ

dt

=

dr

dt

u

r

+r

dθ

dt

u

θ

.

A second diﬀerentiation gives the acceleration

a =

¸

d

2

r

dt

2

−r

dθ

dt

2

¸

u

r

+

¸

r

−1

d

r

2

dθ/dt

dt

¸

u

θ

. (A.4)

Since the acceleration is radial, the coeﬃcient of u

θ

in Eq. (A.4) is zero,

i.e., r

2

dθ/dt = A, a constant. (This is Kepler’s law of areas, i.e., conservation

of angular momentum. Cf. Eq. (3.15).) Set u = 1/r. Then

d

2

r

dt

2

=

d

dθ

dr

dt

dθ

dt

=

d

dθ

dr

du

du

dθ

dθ

dt

dθ

dt

=

d

dθ

−u

−2

du

dθ

Au

2

Au

2

= −A

2

u

2

d

2

u

dθ

2

.

From Eq. (2.1), the coeﬃcient of u

r

in Eq. (A.4) is −κMu

2

. Thus

−κMu

2

=

d

2

r

dt

2

−r

dθ

dt

2

−A

2

u

2

d

2

u

dθ

2

−u

−1

Au

2

2

= −A

2

u

2

d

2

u

dθ

2

+u

or

d

2

u

dθ

2

+u =

κM

A

2

. (A.5)

The solution of this diﬀerential equation is

u =

κM

A

2

[1 +e cos (θ −θ

p

)] ,

where e and θ

p

are constants. (Cf. Eq. (3.20).) This is the polar equation of an

ellipse with eccentricity e and perihelion (point of closest approach to the Sun)

at θ = θ

p

.

89

A.9 The Geodesic Equations

A.9 The Geodesic Equations. This appendix translates the local form

of the geodesic equations Eq. (2.16) to the global form Eq. (2.18). Let (f

ij

)

represent the metric with respect to a local planar frame at P with coordinates

(x

i

). From Eq. (2.6), (f

ij

(P)) = f

◦

, and from the assumption Eq. (2.19),

∂

k

f

ij

(P) = 0.

Let (y

i

) be a coordinate system with metric g(y

i

). We use the notation

y

i

r

=

∂y

i

∂x

r

, x

u

v

=

∂x

u

∂y

v

, y

i

rs

=

∂

2

y

i

∂x

s

∂x

r

, x

u

vw

=

∂

2

x

u

∂y

w

∂y

v

.

Diﬀerentiate ˙ y

i

= y

i

q

˙ x

q

and use Eq. (2.16):

¨ y

i

= y

i

q

¨ x

q

+y

i

qr

˙ x

r

˙ x

q

= y

i

qr

x

r

j

˙ y

j

x

q

k

˙ y

k

.

In terms of components, the formula g

−1

= a

−1

f

◦−1

a

−1

t

from Exercise

2.5c is

g

im

= y

i

r

f

◦ rs

y

m

s

.

Eq. (2.12) shows that Eq. (2.9) holds throughout the local planar frame, not

just at P. Apply ∂

i

to Eq. (2.9), and use Eq. (2.19) and the symmetry of f and

g. Evaluate at P:

∂

i

g

jk

= (∂

p

f

mn

) x

p

i

x

m

j

x

n

k

+f

mn

x

m

ji

x

n

k

+f

mn

x

m

j

x

n

ki

= 2f

◦

mn

x

m

j

x

n

kj

.

Note that (x

u

m

y

m

s

) = (∂x

u

/∂x

s

) is the identity matrix. Substitute the last two

displayed equations in the deﬁnition Eq. (2.17) of the Christoﬀel symbols:

Γ

i

jk

=

1

2

g

im

(∂

k

g

jm

+∂

j

g

mk

−∂

m

g

jk

)

= y

i

r

f

◦ rs

y

m

s

f

◦

uv

x

u

j

x

v

mk

+f

◦

uv

x

u

m

x

v

kj

−f

◦

uv

x

u

j

x

v

km

= y

i

r

f

◦ rs

y

m

s

f

◦

uv

x

u

m

x

v

kj

= y

i

r

f

◦ rs

f

◦

uv

x

u

m

y

m

s

x

v

kj

= y

i

r

f

◦ rs

f

◦

sv

x

v

kj

= y

i

r

x

r

kj

.

Substitute for ¨ y

i

and Γ

i

jk

from above to give the global form Eq. (2.18) of

the geodesic equations:

¨ y

i

+ Γ

i

jk

˙ y

j

˙ y

k

=

y

i

qr

x

q

k

x

r

j

+y

i

r

x

r

kj

˙ y

j

˙ y

k

= ∂

k

y

i

r

x

r

j

˙ y

j

˙ y

k

= 0 .

90

A.10 Geodesic Coordinates

A.10 Geodesic Coordinates. By deﬁnition, the metric f (x) = (f

mn

(x))

of a local planar frame at P satisﬁes (f

mn

(P)) = f

◦

. Given a local planar frame

with coordinates x, we construct a new local planar frame with coordinates ¯ x in

which the metric

¯

f

mn

(¯ x)

additionally satisﬁes ∂

i

¯

f

mn

(P) = 0. The coordinates

are called geodesic coordinates.

We ﬁrst express the derivatives of the metric in terms of the Christoﬀel

symbols by “inverting” the deﬁnition Eq. (2.17) of the latter in terms of the

former. Multiply Eq. (2.17) by (g

ji

) and use the fact that (g

ji

g

im

) = (δ

m

j

),

the identity matrix, to obtain g

ji

Γ

i

kp

=

1

2

[∂

p

g

kj

+∂

k

g

jp

−∂

j

g

kp

]. Exchange j

and k to obtain g

ki

Γ

i

jp

=

1

2

[∂

p

g

jk

+∂

j

g

kp

−∂

k

g

jp

]. Add the two equations to

obtain the desired expression:

∂

p

g

jk

= g

ji

Γ

i

kp

+g

ki

Γ

i

jp

. (A.6)

Deﬁne new coordinates ¯ x by

x

m

= ¯ x

m

−

1

2

Γ

m

st

¯ x

s

¯ x

t

,

where the Christoﬀel symbols are those of the x-coordinates at P. Diﬀerentiate

this and use Γ

m

it

= Γ

m

ti

:

∂x

m

∂¯ x

j

=

∂¯ x

m

∂¯ x

j

−

1

2

Γ

m

sj

¯ x

s

−

1

2

Γ

m

jt

¯ x

t

= δ

m

j

−Γ

m

sj

¯ x

s

.

Substitute this into Eq. (2.9):

¯

f

jk

(¯ x) = f

jk

(x) −f

jn

(x)Γ

n

sk

¯ x

s

−f

mk

(x)Γ

m

sj

¯ x

s

+ 2

nd

order terms.

This shows already that the ¯ x are inertial frame coordinates:

¯

f (P) = f (P) = f

0

.

Diﬀerentiate the above equation with respect to ¯ x

p

, evaluate at P, and apply

Eq. (A.6) to f :

∂

p

¯

f

jk

= ∂

s

f

jk

(P)

∂x

s

∂¯ x

p

−

f

jn

(P)Γ

n

pk

+f

mk

(P)Γ

m

pj

= ∂

p

f

jk

(P) −

f

jn

(P)Γ

n

pk

+f

mk

(P)Γ

m

pj

= 0 .

Thus ¯ x is a geodesic coordinate system.

91

A.11 The Form of the Field Equation

A.11 The Form of the Field Equation. This appendix justiﬁes the

passage from Eq. (2.33) to Eq. (2.34). We assume that Eq. (2.33) holds at E

in local inertial frames at E. From Exercise 2.16, the right side of the equation

transforms contravariantly. Thus if we transform G

jk

contravariantly to other

coordinate systems, then Eq. (2.33) is valid in all coordinate systems:

G

jk

= −8πκT

jk

. (A.7)

Is this equation Eq. (A.7) the form of the ﬁeld equation in all coordinate

systems? First of all, note that the expression of a physical law in one coordinate

system does not uniquely determine its expression in another, just as an equation

of a unit circle in cartesian coordinates (say x

2

+ y

2

= 1) does not uniquely

determine its equation in polar coordinates (r = 1 or r(r + 1) = r + 1 or . . . ).

However, the validity of Eq. (A.7) in local inertial frames implies its validity in

all coordinate systems. Thus there is no loss of generality in taking Eq. (A.7)

as the form of the ﬁeld equation in all coordinate systems.

In the text we assumed that the G

jk

depend only on the g

jk

and their

ﬁrst and second derivatives. We also assumed that ∂T

jk

(E)/∂x

j

= 0 in local

inertial frames at E. From Eq. (A.7), ∂G

jk

(E)/∂x

j

= 0. According to the

mathematical theorem of Lovelock, G

jk

must be of the form given in the left

side of Eq. (2.34).

92

A.12 Eliminating the dtdr Term

A.12 Eliminating the dtdr Term. We make a coordinate change in

Eq. (3.2) to eliminate the dtdr term. From the theory of diﬀerential equations

there is an integrating factor I(t, r) and then a function τ(t, r) so that

dτ = I(e

2µ

dt −udr).

Then

dt = e

−2µ

(I

−1

dτ +udr).

Substitute for dt and then for dτ in the dτdr term:

e

2µ

dt

2

−2udtdr −e

2ν

dr

2

= I

−2

e

−2µ

dτ

2

−(2u

2

e

−2µ

+e

2ν

) dr

2

≡ e

2¯ µ

dτ

2

−e

2¯ ν

dr

2

.

Drop the bars and rename τ to t and this becomes

e

2µ

dt

2

−e

2ν

dr

2

.

There is no dtdr term.

93

A.13 Approximating an Integral

A.13 Approximating an Integral. Retaining only ﬁrst order terms in

the small quantities m/r and m/r

p

we have

(1 −2m/r)

−2

1 −

1 −2m/r

1 −2m/r

p

(r

p

/r)

2

≈

(1 + 2m/r)

2

1 −(1 −2m/r) (1 + 2m/r

p

) (r

p

/r)

2

≈

1 + 4m/r

1 −(1 −2m/r + 2m/r

p

) (r

p

/r)

2

=

1 + 4m/r

1 −(r

p

/r)

2

1 −

2mr

p

r (r +r

p

)

≈

1 −(r

p

/r)

2

−1

(1 + 4m/r)

1 +

2mr

p

r (r +r

p

)

≈

1 −(r

p

/r)

2

−1

1 + 4m/r +

2mr

p

r (r +r

p

)

.

We now see that Eq. (3.25) can be replaced with

t =

rp

re

1 −

r

p

r

2

−1

1 +

4m

r

+

2mr

p

r (r +r

p

)

dr .

Integrate to give

t =

r

2

e

−r

2

p

1

2

+ 2mln

r

e

+

r

2

e

−r

2

p

1

2

r

p

+m

r

e

−r

p

r

e

+r

p

1

2

.

Use r

p

<r

e

to obtain Eq. (3.26).

94

A.14 The RobertsonWalker Metric

A.14 The Robertson-Walker Metric. This appendix derives the

Robertson-Walker for an isotropic universe, Eq. (4.5).

In Chapter 3 we showed that the metric of a spherically symmetric spacetime

can be put in the form Eq. (3.2):

ds

2

= e

2µ

dt

2

−2udtdr −e

2ν

dr

2

−r

2

e

2λ

d Ω

2

, (A.8)

where the coeﬃcients do not depend on φ or θ.

Since the coordinates are comoving, dr = dφ = dθ = 0 for neighboring events

on the worldline of a galaxy. Thus by Eq. (A.8), ds

2

= e

2µ

dt

2

. But ds = dt,

since both are the time between the events measured by a clock in the galaxy.

Thus e

2µ

= 1.

For light sent radially outward from (t, r) to (t +dt, r +dr) we have ds = 0,

and so

0 = dt

2

−2udtdr −e

2ν

dr

2

.

Light sent radially inward from (t, r +dr) must arrive at (t +dt, r) or else there

would not be isotropy half way between r and r +dr. For this second signal,

0 = dt

2

−2udt(−dr) −e

2ν

(−dr)

2

.

Subtract the two equations to give u = 0. Our metric is now of the form

ds

2

= dt

2

−e

2ν

dr

2

−r

2

e

2λ

d Ω

2

(A.9)

If the universe is expanding isotropically about every galaxy, then it seems

that the spatial part of the metric could depend on t only through a common

factor, say S(t), as in Eq. (4.3). We now demonstrate this.

Consider light sent from a galaxy at (t, r, φ, θ) to a galaxy at (t + dt, r +

dr, φ, θ). From Eq. (A.9), dt

2

= e

2ν

dr

2

. Thus the light travel time between

P(r, φ, θ) and Q(r +dr, φ, θ) is e

ν(t,r)

dr. The ratio of this time to that measured

at another time t

is given by e

ν(t,r)

/e

ν(t

,r)

. Similarly, the ratio of the times

between P and R(r, θ, φ +dφ) is e

λ(t,r)

/e

λ(t

,r)

. By isotropy at P the ratios are

equal:

e

ν(t,r)

e

ν(t

,r)

=

e

λ(t,r)

e

λ(t

,r)

. (A.10)

(For example, if the time from the galaxy at P to the nearby galaxy at Q

doubles from t to t

**, then the time from P in another direction to the nearby
**

galaxy at R must also double.) By homogeneity, the two sides of Eq. (A.10)

cannot depend on r. Fix t

**. Then both members of Eq. (A.10) are a function
**

only of t, say S(t). Set ν(t

, r) = ν(r) to obtain e

ν(t,r)

= S(t) e

ν(r)

. Similarly,

e

λ(t,r)

= S(t) e

λ(r)

. Thus the metric Eq. (A.9) becomes

ds

2

= dt

2

−S

2

(t)

e

2ν

dr

2

+r

2

e

2λ

d Ω

2

. (A.11)

The coordinate change ¯ r = re

λ(r)

puts Eq. (A.11) in the form

ds

2

= dt

2

−S

2

(t) ¦e

2ν

dr

2

+r

2

d Ω

2

¦ = dt

2

−S

2

(t) dσ

2

. (A.12)

95

A.14 The RobertsonWalker Metric

We now determine ν in Eq. (A.12). An application of Eq. (2.23) shows that

the curvature K of the half plane θ = θ

0

in the dσ portion of the metric satisﬁes

ν

e

−2ν

= Kr.

By homogeneity, K is independent of r. Integrating thus gives e

−2ν

=

C −Kr

2

, where C is a constant of integration.

Fig. A.12: Evaluating the in-

tegration constant C.

It only remains to determine C. From Eq.

(A.12) we can label the sides of an inﬁnitesimal

sector of an inﬁnitesimal circle centered at the ori-

gin in the φ = π/2 plane. See Fig. A.12. Since the

length of the subtended arc is the radius times the

subtended angle, e

ν(0)

= 1. Thus C = 1.

Set k = 1, 0, −1 according as K > 0, K =

0, K < 0. Then k indicates the sign of the cur-

vature of space. If K = 0, substitute ¯ r = r [K[

1

2

and

¯

S = S [K[

1

2

. Dropping the bars, we obtain

ds

2

= dt

2

−S

2

(t)

dr

2

1 −k r

2

+r

2

d Ω

2

(k = 0, ±1).

The substitution r = ¯ r/(1 +k¯ r

2

/4) gives the Robertson-Walker metric.

96

A.15 The Luminosity-Redshift Relation

A.15 The Luminosity-Redshift Relation. To begin, replace the ap-

proximate relation Eq. (4.2) with this exact relation:

=

L

[ 4πr

2

e

S

2

0

] (1 +z)

2

. (A.13)

According to Eq (4.7) with k = 0 , the term in brackets is the area of the

sphere with coordinate radius r

e

. The 1 +z factors are most easily understood

using the photon description of light. The energy of a photon is E = hf (an

equation discovered by Einstein), where h is Planck’s constant. According to

Eq. (1.7), the energy of each received photon is diminished by a factor 1 + z.

And according to Eq. (1.6), the rate at which photons are received is diminished

by the same factor.

We express r

e

in Eq. (A.13) in terms of z to obtain the luminosity-redshift

relation. Recall that the critical density ρ

c

= 3H

2

/8πκ. Deﬁne

Ω

0

=

ρ

ρ

c

0

=

8πκ

3H

2

0

ρ

0

,

where a “0” subscript means that the quantity is to be evaluated today. From

Eqs. (4.18) and (4.12), 8πκρ/3 = H

2

0

Ω

0

(z+1)

3

. And from Eq. (4.17), Λ/3H

2

0

=

1 −Ω

0

. Substituting in Eq. (4.15) gives

S

S

2

= H

2

0

(1 −Ω

0

) + Ω

0

(z + 1)

3

. (A.14)

From Eq. (4.12), S = S

0

(z + 1)

−1

. Thus S

/S = −(z + 1)

−1

z

. Also, from

Eq. (4.11) with k = 0 , z

= (dz/dr)(dr/dt) = (dz/dr)S

−1

. Substituting into

Eq. (A.14),

dz

dr

2

= S

2

0

H

2

0

(1 −Ω

0

) + (z + 1)

3

Ω

0

. (A.15)

Separating variables, integrating, and substituting into Eq. (A.13) gives the

luninosity-redshift relation:

(z) =

LH

2

0

4π (1 +z)

2

z

0

[(1 −Ω

0

) + (ζ + 1)

3

Ω

0

]

−

1

2

dζ

. (A.16)

97

Index

accelerometer, 13, 31

approximating an integral, 94

approximations, 82

balloon analogy, 67

big bang nucleosynthesis, 69, 75, 81

black hole, 51, 55, 64

blackbody, 73

Braginsky, V. B., 32

Brecher, K., 25

Bridgeman, P.W., 17

Brillit-Hall experiment, 24, 87

Christoﬀel symbols, 41

clock, 11

comoving coordinates, 70

contravariant index, 47

coordinate singularity, 64

coordinates, 14

Copernicus, Nicholas, 29

cosmic background radiation

(CBR), 69, 73–75

cosmological constant, 75

cosmology, 67

covariant index, 47

crab nebula, 51

critical density, 75, 78, 97

curvature, 44, 48

curvature scalar, 46

curve, 12

curved spacetime, 11

dark energy, 76–78

dark matter, 76, 78

degenerate electron pressure, 50

deSitter, Willem, 60

deuterium, 75

Dicke, R., 32

dipole anisotropy, 69

double quasar, 59

dust, 46, 75

Eddington, Sir Arthur Stanley, 58

Einstein summation convention, 38

Einstein tensor, 46

Einstein, Albert, 11, 30, 31

eliminating the dtdr term, 93

energy-momentum tensor, 46

ergosphere, 65

event, 11

Fermi normal coordinates, 48

ﬁeld equation, 44, 45, 47, 52, 74

ﬂat spacetime, 11

form of the ﬁeld equation, 92

frame dragging, 61

Galilei, Galileo, 29

Galle, J., 30

Gauss, C. F., 31, 45

general relativity, 11

geodedic eﬀect, 60

geodesic, 26, 27, 55

geodesic coordinates, 42, 48, 90, 91

geodesic equations, 26, 55, 90

geodesic postulate, 13, 26, 27, 31,

32, 41–43

global coordinate postulate, 36, 37

global coordinates, 35, 37

global positioning system (GPS), 34

gravitational binding energy, 32

gravitational delay, 59, 63

gravitational lens, 59, 76

gravitational radiation, 54, 63, 81

98

gravitomagnetic clock eﬀect, 61

gravitomagnetism, 61

Hafele-Keating experiment, 12, 23,

24, 33, 34, 53

Hils-Hall experiment, 24

homogeneous universe, 95

horizon, 64, 65, 78

Hubble’s constant, 67, 72

Hubble’s law, 67, 68, 72

Hubble, Edwin, 68

inertial, 67

inertial force, 13, 31

inertial frame, 13, 14

inertial frame postulate, 13, 14, 17,

31, 37

inertial object, 13, 31

integrated Sachs-Wolfe eﬀect, 78

intrinsic description, 37, 47

isotropic universe, 70

Joos, G., 24, 86

Kennedy-Thorndike experiment,

24, 88

Kepler, Johannes, 29

Kerr metric, 61, 65

LAGEOS satellites, 61

length contraction, 22, 33, 49, 84

Lense-Thirring eﬀect, 61

LeVerrier, U., 30

light deﬂection, 58

light second, 19

light speed, 25, 40

lightlike separated, 18

local inertial frame, 36

local planar frame, 35, 42

loop quantum gravity, 81

Lovelock’s theorem, 49, 92

luminosity-redshift relation, 77, 97

lunar laser experiment, 32, 60

Macek-Davis experiment, 17, 83

manifold, 36

mass, 48

Mercury, 58

metric postulate, 13, 18, 19, 22, 23,

31, 38–40

Michelson interferometer, 86, 88

Michelson-Morley experiment, 24,

86

Minkowski, Hermann, 19

moving rods, 84

muon, 24

neutron star, 51

Newton, Isaac, 29

newtonian gravitational constant,

29, 53

newtonian orbits, 89

Painlev`e coordinates, 64

periastron advance, 63

perihelion advance, 58

photon, 97

physical constants, 82

planar frame, 14

planar frame postulate, 14, 35

Planck’s constant, 97

Planck’s law, 73

point, 11

postulates, global form, 37

postulates, local form, 37

Pound, R. V., 33

principle of relativity, 84

proper distance, 18

proper time, 40

pseudosphere, 45

Ptolemy, Claudius, 28

pulsar, 51, 63

quantum theory, 81

quasar, 51, 59, 60, 66

Rebka, G. A., 33

redshift, 16

Doppler, 17, 21, 25, 33

expansion, 17, 68, 72

gravitational, 17, 33, 54, 63

Ricci tensor, 46

Riemann, G. B., 37

99

Robertson-Walker metric, 71, 74, 95

rod, 11

round trip light speed, 83

Schwartzschild metric, 53

Schwartzschild radius, 64

Schwartzschild, Karl, 53

Sirius, 50, 54

Sloan Digital Sky Survey (SDSS),

75

Snider, J.L., 33

spacelike separated, 18

spacetime, 11, 19

spacetime coordinates, 14

spacetime diagram, 15

special relativity, 11

standard model, 81

static limit, 65

stellar evolution, 50

STEP, 32

string theory, 81

supernova, 50, 51, 66, 69, 73, 77

surface, 11

surface dwellers, 13

synchronization, 14, 15, 17

tensors, 47

terrestrial redshift experiment, 33,

54

tidal accelerations, 36

time dilation, 20, 24, 33

timelike separated, 18

universal time, 12, 30

vacuum ﬁeld equation, 47

white dwarf, 50

WMAP, 75

worldline, 12

100

To Ellen

.

“The magic of this theory will hardly fail to impose itself on anybody who has truly understood it. 1915 “The foundation of general relativity appeared to me then [1915]. the more transparent his logic. an indiﬀerence occasionally amounting to hostility.” Albert Einstein. It is plain that the clearer the teacher. but not simpler.” Albert Einstein . 1955 “One of the principal objects of theoretical research in any department of knowledge is to ﬁnd the point of view from which the subject appears in its greatest simplicity. the most amazing combination of philosophical penetration.” Max Born. the faster will the pupil learn and the surer and sounder will be his grasp of the subject. physical intuition. and it still does.” Sir Hermann Bondi “Things should be made as simple as possible. I am concerned with the eﬀects of our neglect of foundations on the education of scientists.” Josiah Willard Gibbs “There is a widespread indiﬀerence to attempts to put accepted theories on better logical foundations and to clarify their experimental basis. the greatest feat of human thinking about Nature. the fewer and more decisive the number of experiments to be examined in detail. and mathematical skill.

.

Tensors are not used. This is not a new idea. I believe that this approach provides easier access to and deeper understanding of relativity. independently of any coordinate system. In geometry. independently of any coordinate system. a few basic facts about partial derivatives and line integrals. Then the Pythagorean theorem is proved. The prerequisites are single variable calculus. Similarly. In particular. not computational prowess. its analog for surfaces is presented. it is a serious introduction to the physics and mathematics of general relativity which demands careful study. Before introducing a spacetime concept. but it is used here more systematically than elsewhere. The postulates are in one-to-one correspondence with the fundamental concepts of Riemannian geometry: manifold.Preface My purpose here is to provide. We then prove that there is a metric gjk so that ds2 = gjk (y i ) dy j dy k . and some basic physics. where it belongs. which is similar to the standard approach to the metric in Euclidean geometry. We take a nonstandard approach to the metric of relativity.) Chapter 3 solves the ﬁeld equation for a spherically symmetric spacetime . The theorem is important. metric. and curvature. but the geometric notion of distance is fundamental. with a minimum of mathematical machinery and in the fewest possible pages. geodesic. The basic concepts of Riemannian geometry are developed in order to express these principles mathematically as postulates. the reader has already seen a metric in three other contexts. Chapter 2 introduces the physical principles on which general relativity is based. For example. The approach was chosen to facilitate the transition to general relativity in Chapter 2. (Similarly. and to exhibit clearly the relationship to special relativity and the analogy with surfaces. The ﬁrst two chapters systematically exploit the mathematical analogy which led to general relativity: a curved spacetime is to a ﬂat spacetime as a curved surface is to a ﬂat surface. modern elementary diﬀerential geometry texts often develop the intrinsic geometry of curved surfaces by focusing on the geometric meaning of the metric. The book can stand alone as an introduction to general relativity or it can be used as an adjunct to standard texts. and greatly simpliﬁes the development of the theory. This puts the physical meaning of the metric front and center. Despite it’s brevity and modest prerequisites. distance is ﬁrst understood geometrically. when the metric ds of general relativity is introduced in Chapter 2. The book is for those seeking a conceptual understanding of the theory. a little matrix algebra. The purpose of the postulates is not to achieve complete rigor – which is neither desirable nor possible in a book at this level – but to state clearly the physical principles. Chapter 1 is a self-contained introduction to those parts of special relativity we require for general relativity. we deﬁne the metric of spacial and general relativity physically. a clear and careful explanation of the physical principles and applications of classical general relativity. tensors are not used.

Some 50 exercises are scattered throughout. as elsewhere. but they may be omitted without loss of continuity. giving the new “standard model” of the universe with dark matter and dark energy. We obtain the RobertsonWalker metric in an elementary manner without using the ﬁeld equation. Appendix 1 gives the values of various physical constants. They are best carried out with a computer algebra system. they should be read. I have tried to provide the cleanest possible calculations. In this chapter. There have been many spectacular astronomical discoveries and observations since 1960 which are relevant to general relativity. The chapter closes with short sections on the binary pulsar and black holes. Chapter 4 applies general relativity to cosmology. Appendix 2 contains several approximation formulas used in the text. Some material has been placed in 14 appendices to keep the main line of development visible. The appendices occasionally require more background of the reader than the text. We then solve the ﬁeld equation with a nonzero cosmological constant for a ﬂat Robertson-Walker spacetime. The geodesic equations are then solved and applied to the classical solar system tests of general relativity. There is a discussion of the Kerr metric. They often serve as examples of concepts introduced in the text. Some tedious (but always straightforward) calculations have been omitted.to obtain the Schwartzschild metric. including gravitomagnetism and its observation by the LAGEOS satellites. . We describe them at appropriate places in the book. WMAP data allows us to determine all unknown parameters in the solution. If they are not done.

. . . .2 The Schwartzschild Metric . . . . . . . . . . 64 Cosmological Spacetimes 4. . . . . . . . . . . . .6 The Field Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 The Local Inertial Frame Postulate . . . . . . . . . . . . . . . . . . . . . . 26 Curved Spacetimes 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Our Universe II . . . .5 General Relativity Today . . . . . . . . . . . . . . . . . . . . . 18 1. . .Contents Preface Contents 1 Flat Spacetimes 1. . . . . . . 41 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 2. . .1 Spacetimes . . . . . 57 3. . . . . . . . . . .50 3. . . . . . . . . . . . . . . . . . . . . . . . . . . 72 4. . . . . . . . . . . . . 61 3. . . . . . .3 The Metric Postulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 1. . . . . . . . . . . . 80 Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . .1 History of Theories of Gravity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6 Black Holes . . . . . . . 44 Spherically Symmetric Spacetimes 3. . . . . . . . 28 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 The Geodesic Postulate . . . . . . . . . .2 The Inertial Frame Postulate . . . .1 Our Universe I .2 The Key to General Relativity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 2. . . . . 63 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Stellar Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 The Solar System Tests . . . . . . . 74 4. . .2 Robertson-Walker Spacetimes . . . . . . . . . . . . . . . . . . . . . . 38 2. . . . . . . . . . . . . . . . . 11 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 Kerr Spacetimes . . . . . . . . . . . . . . . . . . .4 The Geodesic Postulate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5 The Binary Pulsar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 The Metric Postulate . . . . . . . . . . . . . . .67 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 The Expansion Redshift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Index 2 3 4 . . . . .

.

An event is a speciﬁc time and place. In relativity the fundamental entities are events. To attend the concert. This analogy will be a major theme of this book. A curved spacetime is one with signiﬁcant gravity. We shall explore spacetimes with clocks to measure time and rods (rulers) to measure space. Albert Einstein created the theory during the decade following the publication. a spacetime is a set of events. There is nothing mysterious about the words “ﬂat” or “curved” attached to a set of events.Chapter 1 Flat Spacetimes 1.) Similarly. i. A ﬂat spacetime is one without signiﬁcant gravity. (We shall prefer the term “ﬂat surface” to “plane”. For example. in 1905. we shall use the mathematics of ﬂat and curved surfaces to guide our understanding of the mathematics of ﬂat and curved spacetimes. The special theory is a theory of space and time which applies when gravity is insigniﬁcant. A point is a speciﬁc place. and gravity. already hinted at. we might consider the events in a speciﬁc room between two speciﬁc times.e. the collision of two particles is an event. In geometry the fundamental entities are points. It is commonly felt to be the most beautiful of all physical theories. time. concerning the mathematical description of a spacetime: a curved spacetime is to a ﬂat spacetime as a curved surface is to a ﬂat surface. of his special theory of relativity. distance. A ﬂat or curved surface is a set of points. Special relativity describes ﬂat spacetimes. General relativity describes curved spacetimes. you must be at the time and the place of the event. clocks and rods do not in fact live up to all that we usually expect of them. The general theory generalizes the special theory to include gravity. However.1 Spacetimes The general theory of relativity is our best theory of space. A concert is an event (idealizing it to a single time and place).. For example. 11 . In this section we shall see what we expect of them in relativity. They are chosen because of a remarkable analogy.

1.1 Spacetimes Clocks. First a deﬁnition. A curve is a continuous succession of points in a surface. Similarly, a worldline is a continuous succession of events in a spacetime. A moving particle or a pulse of light emitted in a single direction is present at a continuous succession of events, its worldline. Even if a particle is at rest, time passes, and the particle has a worldline. The length of a curve between two given points depends on the curve. Similarly, the time between two given events measured by a clock moving between the events depends on the clock’s worldline! J. C. Hafele and R. Keating provided the most direct veriﬁcation of this in 1971. They brought two atomic clocks together, placed them in separate airplanes which circled the Earth in opposite directions, and brought the clocks together again. Both clocks moved between the event of their separation and the event of their reunion. The clocks measured diﬀerent times between the events. The diﬀerence was small, about 10−7 sec, but was well within the ability of the clocks to measure. There is no doubt that the eﬀect is real. Relativity predicts the measured diﬀerence. Exercise 1.9 shows that special relativity predicts a diﬀerence between the clocks. Exercise 2.1 shows that general relativity predicts a further diﬀerence. Exercise 3.3 shows that general relativity predicts the observed diﬀerence. Relativity prtedicts large diﬀerences between clocks whose relative velocity is close to that of light. The best answer to the question “How can the clocks in the experiment possibly disagree?” is the question “Why should they agree?” After all, the clocks are not connected. According to everyday ideas they should agree because there is a universal time, common to all observers. It is the duty of clocks to report this time. The concept of a universal time was abstracted from experience with small speeds (compared to that of light) and clocks of ordinary accuracy, where it is nearly valid. The concept permeates our daily lives; there are clocks everywhere telling us the time. However, the Hafele-Keating experiment shows that there is no universal time. Time, like distance, is route dependent. Since clocks on diﬀerent worldlines between two events measure diﬀerent times between the events, we cannot speak of the time between two events. However, the relative rates of processes - the ticking of a clock, the frequency of a tuning fork, the aging of an organism, etc. - are the same along all worldlines. (Unless some adverse physical condition aﬀects a rate.) Twins traveling in the two airplanes of the Hafele-Keating experiment would each age according to the clock in their airplane. They would thus be of slightly diﬀerent ages when reunited.

12

1.1 Spacetimes Rods. We need another deﬁnition. Consider astronauts in interstellar space, where gravity is insigniﬁcant. If their rocket is not ﬁring and their ship is not spinning, then they will feel no forces acting on them and they can ﬂoat freely in their cabin. If their spaceship is accelerating, then they will feel a force pushing them back against their seat. If the ship turns to the left, then they will feel a force to the right. If the ship is spinning, they will feel a force outward from the axis of spin. Call these forces inertial forces. Accelerometers measure inertial forces. Fig. 1.1 shows a simple accelerometer that detects any motion of a weight held at the center of a frame by springs. An inertial object experiences no inertial forces. If an object is inertial, then any object moving at a constant velocity with respect to it is also inertial. In special relativity we make an assumption which allows us to speak of the distance between two inertial objects at rest with respect to each other: Inertial rigid rods side by side and at rest with respect to two inertial objects measure the same distance between the objects. Fig. 1.1: An acceleromMore precisely, we assume that any diﬀerence is due to eter. The weight is some adverse physical cause (e.g., thermal expansion) held at the center by to which an “ideal” rigid rod would not be subject. In springs. Acceleration particular, the history of a rigid rod does not aﬀect its causes the weight to move from the center. length. Noninertial rods are diﬃcult to deal with in relativity, and we shall not consider them. In the next three sections we give three postulates for special relativity. The inertial frame postulate asserts that certain natural coordinate systems, called inertial frames, exist for a ﬂat spacetime. The metric postulate asserts a universal light speed and a slowing of clocks moving in inertial frames. The geodesic postulate asserts that inertial particles and light move in a straight line at constant speed in inertial frames. We shall use the analogy mentioned above to help us understand the postulates. Imagine two dimensional beings living in a ﬂat surface. These surface dwellers can no more imagine leaving their two spatial dimensions than we can imagine leaving our three spatial dimensions. Before introducing a postulate for a ﬂat spacetime, we introduce the analogous postulate formulated by surface dwellers for a ﬂat surface. The postulates for a ﬂat spacetime use a time dimension, but those for a ﬂat surface do not.

13

1.2 The Inertial Frame Postulate

1.2

The Inertial Frame Postulate

Surface dwellers ﬁnd it useful to label the points of their ﬂat surface with coordinates. They construct, using identical rigid rods, a square grid and assign rectangular coordinates (x, y) to the nodes of the grid in the usual way. See Fig. 1.2. They specify a point by using the coordinates of the node nearest the point. If more accurate coordinates are required, they build a ﬁner grid. Surface dwellers call the coordinate system a planar frame. They postulate:

**The planar frame postulate for a ﬂat surface
**

A planar frame can be constructed with any point P as origin and with any orientation. Similarly, it is useful to label the events in a ﬂat spacetime with coordinates (t, x, y, z). The coordinates specify when and where the event occurs, i.e., they completely specify the event. We now describe how to attach coordinates to events. The procedure is idealized, but it gives a clear physical meaning to the coordinates. To specify where an event occurs, construct, using identical rigid rods, an inertial cubical lattice. See Fig. 1.3. Assign rectangular coordinates (x, y, z) to the nodes of the lattice in the usual way. Specify Fig. 1.2: A planar frame. where an event occurs by using the coordinates of the node nearest the event. To specify when an event occurs, place a clock at each node of the lattice. Then the times of events at a given node can be speciﬁed by reading the clock at that node. But to compare meaningfully the times of events at diﬀerent nodes, the clocks must be in some sense synchronized. As we shall see soon, this is not a trivial matter. (Remember, there is no universal time.) For now, assume that the clocks have been synchronized. Then specify when an event occurs by using the time, t, on the clock at the node nearest the event. And measure the coordinate time diﬀerence Fig. 1.3: An inertial lat- ∆t between two events using the synchronized clocks tice. at the nodes where the events occur. Note that this requires two clocks. The four dimensional coordinate system obtained in this way from an inertial cubical lattice with synchronized clocks is called an inertial frame. The event (t, x, y, z) = (0, 0, 0, 0) is the origin of the inertial frame. We postulate:

**The inertial frame postulate for a ﬂat spacetime
**

An inertial frame can be constructed with any event E as origin, with any orientation, and with any inertial object at E at rest in it. 14

Say that the clocks are synchronized if tQ = tP + D/v. (1. Let it arrive at a node Q of the lattice at time tQ according to the clock at Q. 1.2) i. and call them synchronized? We might try the following deﬁnition. Fig. Why not simply bring the clocks together. This deﬁnition is ﬂawed because it contains a logical circle: v is deﬁned by a rearrangement of Eq. We adopt the following deﬁnition.4 shows the worldlines of two particles. Fig. Synchronized clocks cannot be deﬁned using v because synchronized clocks are needed to deﬁne v. Let it arrive at a node Q at time tQ according to the clock at Q. First of all.4: Two worldlines. it must be answered with a deﬁnition. We return to the matter of synchronizing the clocks in the lattice.2. 1. A reformulation of the deﬁnition will make it more transparent.2 The Inertial Frame Postulate If we suppress one or two of the spatial coordinates of an inertial frame. move them to the nodes of the lattice. essentially due to Einstein.1): v = D/(tQ − tP ). Exercise 1. One is at rest on the x-axis and one moves away from x = 0 and then returns more slowly.5.e. 1. For example. Emit a pulse of light from a node P at time tP according to the clock at P . rather. Fig. if the times in the two directions are equal. (1. Similarly.1) Intuitively.. the term D/v compensates for the time it takes the signal to get to Q.1.1. Send a signal from a node P of the lattice at time tP according to the clock at P . The clocks are synchronized if tQ − t P = tP − tQ . 1. Show that the worldline of an object moving with constant speed v along the x-axis is a straight line with slope v. synchronize them. emit a pulse of light from Q at time tQ and let it arrive at P at tP . Exercise 1. Let the distance between the nodes be D and the speed of the signal be v. (1. then we can draw a spacetime diagram and depict worldlines. what does it mean to say that separated clocks are synchronized? It was Einstein who ﬁrst realized that the answer to this question is not given to us by Nature.5: Worldline of a particle moving with constant speed. Then tQ = tQ 15 . See Fig. Suppose the pulse from Q to P is the reﬂection of the pulse from P to Q.

fo (1.3) is independent of the time the pulse is sent. (We use ∆s rather than ∆t to conform to notation used later in more general situations. (1. T in Eq. There is a tacit assumption in the deﬁnition of synchronized clocks that the two sides of Eq.5) where ∆so = tS − tS is the time between the observation of the pulses at S and ∆se = tR − tR is the time between the emission of the pulses at R.1) is not.6) are successive wavecrests of light emitted at frequency fe = (∆se )−1 and observed at frequency fo = (∆so )−1 . If. Exercise 1.3) is a satisfactory deﬁnition but Eq. then Eq.3. Let them arrive at a node S at times tS and tS according to a clock at S. If the two “pulses” of light in Eq.6) Equations (1. then the observer at S would see clocks at R. for example. Substitute this in Eq. Show that with the assumption Eq. (1. (1. (1. A rearrangement of Eq.6) can be written z= fe − 1. then Eq.5) states that an observer at S sees (actually sees) the clock at R going at the same rate as his clock. Of course.7) 16 .4) gives ∆so = ∆se . z = 1 (∆so /∆se = 2). (1. (1.1.5) correspond to z = 0.2): t Q = tP + T . proceed at half the rate they do at S. Exercise 1. We will encounter situations in which ∆so = ∆se . Explain why Eq. (1.4) and (1. (1. (1. Let 2T be the round trip time: 2T = tP − tP . (1. and all other physical processes at R. (1. (1.4.2 The Inertial Frame Postulate in Eq. the observer at S will see all physical processes at R proceed at the same rate they do at S. Deﬁne the redshift z= ∆so − 1. (1.3) the clocks are synchronized if the pulse arrives at Q in half the time it takes for the round trip.) If a clock at R emits pulses of light at regular intervals to S.4). (1. ∆se (1.2).2) do not depend on the times that the pulses are sent: Emit pulses of light from a node R at times tR and tR according to a clock at R. Then tS − t R = t S − tR .4) With this assumption we can be sure that synchronized clocks will remain synchronized.

(1.6 we shall see that Eq. In an experiment performed in 1963. Let the pulse be at O. if the emitter and observer are in relative motion in a ﬂat spacetime. O at times tO .8) gives tQ − t P = t P − tQ . The inertial frame postulate asserts in part that clocks in an inertial lattice can be synchronized according to the deﬁnition Eq. the clocks at P and Q are synchronized. tQ .6.3) angle in opposite directions. 1. veriﬁed the assumption for a square to one part in 1012 . When the pulse arrives at Q set the clock Fig. or. 4. See Fig. W. Emit a pulse of light at O toward Q at time tO according to the clock at O. W. Similarly. Thus.. To show that the clocks at any two nodes P and Q are now synchronized with each other. The reader may skip the proof and turn to the next section without loss of continuity. We now prove this with the aid of an auxiliary assumption.1. Q. Macek and D. Later we shall see two other kinds of redshift: gravitational redshifts in Sec. z = 0. since the clock at O is synchronized with those at P and Q. P. 2. let the times for a pulse sent around in the other direction be tO .2). Synchronize all clocks with the one at O in this way. Davis.2 The Inertial Frame Postulate In Exercise 1. Jr. (1. Let 2T be the time. 17 .2 and expansion redshifts in Sec. See 3.1. We have the algebraic identities tR − tO = (tR − tP ) + (tP − tQ ) + (tQ − tO ) tR − tO = (tR − tP ) + (tP − tQ ) + (tQ − tO ). (1. See Fig. the clocks at O and Q are now synchronized.6: Light traversing a trithere to tQ = tO + T . tR according to the clock at that node.e. tP . tR . (1.. we can “spread time over space”. Also. Bridgeman’s descriptive phrase. in P. the left sides of the two equations are equal. 1.e. tP . T. as measured by a clock at the origin O of the lattice. M. 1. subtracting the equations Eq.6. (1. According to Eq. M. tQ .8) According to our assumption. we must make this assumption: The time it takes light to traverse a triangle in the lattice is independent of the direction taken around the triangle. i.5) is violated. This is called a Doppler redshift. tP − tO = tR − tP and tR − tQ = tQ − tO . for light to travel from O to another node Q and return after being reﬂected at Q. Reﬂect a pulse of light around the triangle OP Q. i. The three types of redshifts have diﬀerent physical origins and so must be carefully distinguished.

9) will always produce ∆s. Then ∆s2 = ∆x2 + ∆y 2 . An inertial rigid rod can have its ends present at E and F simultaneously.9) . This peculiar deﬁnition will bring uniﬁcation to the three kinds of separation. It is called the in both planar frames. ∆y) in a planar frame. E and F can be on the worldline of an inertial clock. 18 (1. Equivalently. separated: • Lightlike separated. They call ∆s the proper distance between the points. E and F are simultaneous in an inertial frame in which the rod is at rest.7. The two of them together do: they determine ∆s. independently of any planar frame. (1. But surface dwellers single out for special study the distance ∆s along the straight line between P and Q. 1. The proper distance ∆s is deﬁned geometrically. (1. The reason for the absolute value will become clear later. independent of any coordinate system. Let ∆s be the time between E and F measured by an inertial clock.3 The Metric Postulate 1.3 The Metric Postulate Let P and Q be points in a ﬂat surface. there Fig. 1. This is the proper distance between the events. which has a direct geometric signiﬁcance. Given two events in a ﬂat spacetime.9) is a quantity ∆s relating them. Let P and Q have coordinate diﬀerences (∆x. we say. See Fig. This is the proper time between the events.) Let |∆s | be the length the rod. The coordinate diﬀerences ∆x and ∆y between P and Q are diﬀerent in diﬀerent planar frames. But surface dwellers discover a simple formula for ∆s in terms of the coordinate diﬀerences of two points in a planar frame: The metric postulate for a ﬂat surface Let ∆s be the proper distance between points P and Q. E and F can be on the worldline of a pulse of light. Neither ∆x nor ∆y has a geometric signiﬁcance independent of the particular planar frame chosen. (Simultaneously means that light pulses emitted at E and F reach the center of the rod simultaneously. The deﬁnition of ∆s in a ﬂat spacetime is more complicated than in a ﬂat surface. • Timelike separated. • Spacelike separated.7: ∆s is given by Eq. the particular combination of the diﬀerences in Eq. as there are three ways in which events E and F can be. Diﬀerent curves between the points have diﬀerent lengths. (spacetime) interval between the events. But we single out for special study the proper time ∆s . Set ∆s = 0 for E and F . However.1. Other clocks moving between the events will measure diﬀerent times.

Since by deﬁnition ∆s = 0 for lightlike separated events. Eq. (1. Call this one (light) second of distance.1. are doomed to fade away into mere shadows. are diﬀerent in diﬀerent inertial frames. Eq. a universal light speed means c = |∆x/∆t| = 1 . in the words of Hermann Minkowski. This shows that the joining of space and time into spacetime is not an artiﬁcial technical trick. i. ∆z) in an inertial frame. The four of them together do: they determine ∆s. (1. and all other speeds are expressed as a fraction of the speed of light. (1. ∆x. (1. With c = 1.10) The coordinate diﬀerences between E and F . 19 . Ordinarily the fractions are very small.the light year. (1. ∆t = ∆s and ∆x = ∆y = ∆z = 0. It will be convenient to drop the y.10) will always produce ∆s. Then ∆s2 = ∆t2 − ∆x2 − ∆y 2 − ∆z 2 . “Space by itself. This will not be the case in an inertial frame in which the clock is moving. independently of any inertial frame. and time by itself. But there is a simple formula for ∆s in terms of the coordinate diﬀerences of two events in an inertial frame: The metric postulate for a ﬂat spacetime Let ∆s be the interval between events E and F . For example. including the time coordinate diﬀerence. In an inertial frame in which the clock is at rest. Rather. ∆y.11) Lightlike separated events. and only a kind of union of the two will preserve an independent reality. We show that a universal light speed implies the metric postulate. independent of any inertial frame. However.. Let the events have coordinate diﬀerences (∆t. (You are probably familiar with a similar unit of distance . Choose the second as the unit of time.3 × 10−11 sec. No one of the coordinate diﬀerences has a physical signiﬁcance independent of the particular inertial frame chosen.) Then 1 cm = 3.10) and use ∆s2 = ∆t2 − ∆x2 . With this convention c = 1.” The metric postulate is based on the fundamental assumption of a universal light speed: the speed of light has always the same value c in all inertial frames.and z-coordiantes in Eq. |∆x| = |∆t|. who introduced the spacetime concept in 1908. the particular combination of the diﬀerences in Eq.3 The Metric Postulate The spacetime interval ∆s is deﬁned physically.11) is satisﬁed. In this case E and F can be on the worldline of a pulse of light. suppose that an inertial clock measures a proper time ∆s between two events. Then choose the distance light travels in one second – about 3 × 1010 cm – as the unit of distance. which has a direct physical signiﬁcance.e.11). (1. With a suitable choice of units of time and space we can choose c = 1 .

the light takes the time ∆t from E to M to F . Then from Eq. Since c = 1 in I. In I . 1 the length of the rod is 2 ∆s in I . Let C carry a rod R perpendicular to its direction of motion. In I.3 The Metric Postulate Timelike separated events. since the light travels farther in I than in I (the hypotenuse twice vs. The argument shows how it is possible for a single pulse of light to have the same speed in inertial frames moving with respect to each other: the speed (distance/time) can be the same because the distance and the time are diﬀerent in the two frames.12) the time 20 . 1. (1.11) is satisﬁed. By deﬁnition. e and f are the points at which events E Refer to the rightmost trian. This is a contradiction.5. the light travels a distance ∆t in I.1. We show that Eq. 1. (1.8. Criticize the following argument. (1. We can express the proper time ∆s in a diﬀerent way. Informally. The length of R is chosen so that the pulse is reﬂected by M back to F . C is at rest in some inertial frame I .and F occur. This gives the labeling of the hypotenuse. Appendix 4 discusses this. In this case E and F can be on the worldline of an inertial clock C. Since c = 1 in I . In short. R. This gives the labeling of the base of the triangle. This gives the labeling of the altitude of the triangle. the light travels the length of the rod twice in the proper time ∆s between E and F measured by C. “moving clocks run slowly”. 1. This shows. (There is a tacit assumption here that the length of R is the same in I and I .12) The proper time between two events is less than the time determined by the synchronized clocks of an inertial frame: ∆s < ∆t. This is called time dilation. But exactly the same argument carried out in I will show that the time between the events is greater in I than in I. the proper time is ∆s = (∆t2 − ∆x2 ) 2 = [1 − (∆x/∆t)2 ] 2 ∆t = (1 − v 2 ) 2 ∆t. in a most graphic way. together with C. 1 1 1 (1.11) is satisﬁed for timelike separated events. the distance between E and F is ∆x.8 shows the path of the light in I. the time (= distance/speed) between E and F is longer in I than for C. At E a light pulse is sent along R from C. the altitude twice) and the speed c = 1 is the same in I and I . that accepting a universal light speed forces us to abandon a universal time. gle in Fig.8: ∆s2 = ∆t2 −∆x2 for timelike separated the light reﬂects oﬀ M . (1. In I. events. and M as Fig. Fig. ∆s is the time C measures between the events. We have just seen that the time between two events is greater in I than in I . Exercise 1. Let v = |∆x|/∆t be the speed of an inertial clock. Let R have a mirror M on its end. According to Eq.11).) Applying the Pythagorean theorem to the triangle shows that Eq.

Since c = 1 in I. b. or timelike separated (|∆x| < |∆t|) . v is very small. ∆t/∆s = 1 (1 − v 2 )− 2 → ∞.12). v. as expected.3 The Metric Postulate dilation factor is (1 − v 2 ) 2 . 1. between the 1 timelike separated events S and E.6) by setting ∆s = ∆t. But as v → 1. This case can be omitted without loss of continuity. the equations for L+ and W give S(∆x. Show that with c = 1.10: ∆s2 = ∆t2 − ∆x2 for spacelike separated events E and F . If E and F are not lightlike separated (|∆x| = |∆t|) . Let ∆te be the time between the emission of pulses. a. ∆t). Fig. z = [(1 + 1 1 v)/(1 − v)] 2 − 1 . as measured by O . the slope of L± is ±1.10 shows the worldline W of an inertial observer O moving with velocity v = ∆t/∆x in I. Thus |∆t/∆x| < 1 . (1. Spacelike separated events. (That is not a typo. c. 1. and ∆to be the time between the reception of pulses by the observer. then |∆x| > |∆t| . Exercise 1. Since the times are equal. According to Eq. 1 Fig. For convenience. Use the result of part Fig. 1.) L± are the light worldlines through F . Investigate the Doppler redshift. (1. Similarly. Show that ∆to = ∆te + v∆te /c . Let a source of light pulses move with velocity v directly away from an observer at rest in an inertial frame.9 shows the graph of ∆t/∆s vs.11). v 2 is even smaller. is (∆x2 −∆t2 ) 2 .9: ∆t/∆s = (1 − v 2 )− 2 . 21 .1. 1. the proper time. simultaneous in an inertial frame I in which O is at rest. let E have coordinates E(0. Fig. 0). a. See the text. Show that z = v/c in this approximation. For normal speeds. as it plays only a minor role in this book. E and F are. Solve simultaneously the equations for L− and W to obtain the coordinates R(∆x. (1. Ignore time dilation in Eq. and so from Eq. ∆t). and between E and R.6. ∆s ≈ ∆t. by deﬁnition.

7. This proves Eq. Local Forms. a formula for proper time in inertial frames. then ds2 = and the length of the curve is b b dx dp 2 + dy dp 2 dp2 .11) for spacelike separated events. y(p)). local form Let P and Q be neighboring points. A calculation similar to Eq. and a formula for proper distance in inertial frames. y = r sin θ. Calculate the circumference of the circle x = r cos θ. the metric postulate is a mathematical expression of three physical assertions: a universal light speed in inertial frames. Then ds2 = dx2 + dy 2 . which we discuss in Appendix 4. Let the points have coordinate diﬀerences (dx. (This is not length contraction. Thus. dy) in a planar frame. Eq. the distance between E and F in I is (∆x2 − ∆t2 ) 2 . This is the length of a rod at rest in I with its ends simultaneously at E and F .9) gives the distance ds between neighboring points along any curve: The metric postulate for a ﬂat surface. Of course.3 The Metric Postulate Since c = 1 in I .12) gives |∆s | = (1 − v 2 ) 2 |∆x|. diﬀerent curves between two points can have diﬀerent lengths. By deﬁnition. The diﬀerential version of Eq. 22 .13) The proper distance between spacelike separated events is less than an inertial frame distance: |∆s | < |∆x|. 0 ≤ θ ≤ 2π. (1. a ≤ p ≤ b. Exercise 1. gives only the distance along a straight line between two points. 1 1 (1. Let ds be the distance between them. (1. The metric postulate for a planar frame. (1. (1. s= p=a ds = p=a dx dp 2 + dy dp 2 1 2 dp . if a curve is parameterized (x(p). this is the proper distance |∆s | between the events.1.9).) In summary.

9. From Eq. (Thus the clock on the ground is not at rest. (1. Let a clock move between two events with a time diﬀerence 1 ∆t.14) s= p=a ds = p=a dt dp 2 − dx dp 2 − dy dp 2 − dz dp 2 1 2 dp. Exercise 1. as measured by the clock. Let the events have coordinate diﬀerences (dt. The diﬀerential version of Eq. 23 . Assume that the Earth is spinning on its axis at one revolution per 24 hours in an inertial frame.1. let ds be the time between them as measured by any (inertial or noninertial) clock. (1. let ds = 0. Then ds2 = dt2 − dx2 − dy 2 − dz 2 . Consider a simpliﬁed Hafele-Keating experiment.8. is b b (1.3 The Metric Postulate The metric postulate for an inertial frame Eq.4×10−7 s. If E and F are lightlike separated. One clock remains on the ground and the other circles the equator in an airplane to the west . If the events are timelike separated. dz) in an inertial frame.10) gives the time ds measured by any clock between neighboring events on its worldline: The metric postulate for a ﬂat spacetime. 2 a Suppose that ∆t = 40 hours and the speed of the airplane with respect to the ground is 1000 km/hr. local form Let E and F be neighboring events.14). Let v be the small constant speed of the clock.8 for each clock to show that the diﬀerence between the clocks due to time dilation is ∆sa − ∆sg = 1 2 2 (v − vg )∆t. Exercise 1. dx. y(p). z(p)). Show that ∆t−∆s ≈ 2 v 2 ∆t. Substitute values to obtain ∆sa −∆sg = 1.10) is concerned only with times measured by inertial clocks. then the time s to traverse the worldline. a ≤ p ≤ b.) Notation: ∆t is the duration of the trip in the inertial frame. (1. dy. In general. vg is the velocity of the clock on the ground and ∆sg is the time it measures for the trip.opposite to the Earth’s rotation. va and ∆sa are deﬁned similarly for the airplane. if the worldline of a clock is parameterized (t(p). clocks on diﬀerent worldlines between two events will measure different times between the events. Use Exercise 1. x(p).

According to the diﬀerential version of Eq. A. The Michelson-Morley experiment is described in Appendix 5. and so this is a test of the local form Eq. The speed c of light can be measured by sending a pulse of light from a point P to a mirror at a point Q at distance D and measuring the elapsed time 2T for it to return. In a famous experiment. and inertial frames and at diﬀerent times.12). W. Morley compared the two way speed of light in perpendicular directions from a given point.1). Their experiment has been repeated many times. See Appendix 7. (1. during which time the Earth moved to the opposite side of its orbit. performed by R. (1. performed in 1887. L. Then c = 2D/2T . who found that any diﬀerence in the two way speeds is less than six parts in 1012 . most accurately by G. and because the uncertainty of the experiment is large (±10%). Hall. on days six months apart. then its average life. There is excellent evidence for a universal light speed. They thus provide strong motivation for our deﬁnition of synchronized clocks: If the two way speed of light has always the same value.14) of the metric postulate as well as the original form Eq. A modern version of the experiment using lasers was performed in 1979 by A. These experiments provide good evidence that the two way speed of light is the same in diﬀerent directions. Another experiment. if the muon is moving in a circle with constant speed v. then the speed of light from P to Q is equal to the speed from Q to P . giving a time dilation factor ∆t/∆s = (1 − v 2 )− 2 = 29! The circular motion was accompanied by an acceleration of 1021 cm/sec2 . A. J. should be larger by a factor (1 − v 2 )− 2 . A more recent experiment by D. Because general relativistic eﬀects play a part in the Hafele-Keating experiment (see Exercise 2. M. found the two way speed of light to be the same. Brillit and J. places. (1. within six parts in 109 . When at rest the average lifetime of a muon is 3 × 10−6 sec. See Appendix 6.10). Kennedy and E. as measured in the 1 laboratory. Equation 1. First of all. (1. Hils and Hall improved the result by over two orders of magnitude.3 The Metric Postulate Experimental Evidence. In the experiment 1 v = . They found that any diﬀerence in the two way speed of light in perpendicular directions is less than four parts in 1015 . Much better evidence comes from observations of subatomic particles called muons. this experiment is not a precision test of time dilation for clocks. c is a two way speed. Thorndike in 1932. measured with a single clock.2%. An experiment performed in 1977 showed this within an experimental error of . what could be more natural than to deﬁne synchronized clocks by requiring that the one way speed have this value? 24 . We emphasize that with our deﬁnition of synchronized clocks this equality is a matter of deﬁnition which can be neither conﬁrmed nor refuted by experiment.9994. Thus the one way speed of light can be measured by measuring the two way speed. realize that if clocks at P and Q are synchronized according to the deﬁnition Eq. Joos in 1930. Inertial frames in which the Earth is at rest on days six months apart move with a relative speed of 60 km/sec (twice the Earth’s orbital speed).2).5 shows that this two way speed is equal to the one way speed from P to Q.1. Michelson and E.

25 .6. These systems are described in Sec. The universal nature of the speed of light makes possible the modern deﬁnition of the unit of length: “The meter is the length of the path traveled by light during the time interval of 1/299.1.458 of a second. the speed of light is 299. The speed speed of its source. If light were like baseballs.458 m/sec.1.11: The speed of light is independent of the panion directly. of the neutron star toward or away from the Earth can be determined from the Doppler redshift of the time between pulses.Fig.11. 1. X-rays emitted when the neutron star is moving toward the Earth could catch up with those emitted earlier when it was moving away from the Earth. an analysis of the arrival times of the pulses at Earth made in 1977 by K.” Thus. 1.) Finally. and the neutron star could be seen coming and going at the same time! See Fig. Brecher shows that no more than two parts in 109 of the speed of the source is added to the speed of the X-rays. For example. See Exercise 1.3 The Metric Postulate In all of the above experiments. then the speed of a moving source would be imparted to the speed of light it emits.792. recall from above that the universal light speed statement of the metric postulate implies the statements about timelike and spacelike separated events. then various strange eﬀects would be observed. (It is not possible to “see” the neutron star in orbit around its com.792. This does not happen. the source of the light is at rest in the inertial frame in which the light speed is measured. by deﬁnition. 3. Strong evidence that this is not so comes from observations of certain neutron stars which are members of a binary star system and which emit X-ray pulses at regular intervals. If the speed of the star were imparted to the speed of the X-rays. Thus the evidence for a universal light speed is also evidence for the other two statements.

x2 (s) = (sin θ)s + b2 . i = 1.15). 26 . Diﬀerentiate twice with respect to s to obtain The geodesic postulate for a ﬂat surface Parameterize a straight line with areclength s . The line in Fig. 2. 1. Then in every planar frame xi = 0.12 can be parameterized by the (proper) distance s from (b1 . Not all parameterizations of a straight line satisfy the geodesic diﬀerential equations Eq.12: A geodesic in a planar frame.15) The straight lines are called geodesics.4 The Geodesic Postulate We will ﬁnd it convenient to use superscripts to distinguish coordinates. x2 ) instead of (x.4 The Geodesic Postulate 1. 1. b2 ) to (x1 . Surface dwellers ﬁnd that straight lines in their surface can be parameterized xi (s) = ai s + bi . Fig. x2 ) : x1 (s) = (cos θ)s + b1 . 2. Example: xi (p) = ai p3 + bi . ¨ (1. y) for planar frame coordinates. (1. Thus we use (x1 .1. i = 1.

The geodesic postulate is a mathematical expression of our physical assertion that inertial particles and light move in a straight line at constant speed in an inertial frame. (1. 3 (1. Eq. Make as long a list as you can of analogous properties of ﬂat surfaces and ﬂat spacetimes. 2.16) (Diﬀerentiate to give dxi /dx0 = ai . the velocity is constant. Set x0 = p. Exercise 1.4 The Geodesic Postulate We will ﬁnd it convenient to use (x0 . i = 0.11. and b0 = 0. Show the worldline can also parameterized with s.1. Also. The geodesic postulate for a ﬂat spacetime Worldlines of inertial particles and pulses of light can be parameterized so that in every inertial frame xi = 0. the proper time along the worldline. x1 . z) for inertial frame coordinates. 3.17) in an inertial frame. Eq. a parameter. 1. is symmetric in all four coordinates of the inertial frame. unlike Eq. The worldlines are called geodesics.18) 27 .17) parameterizes the worldline of an inertial particle with x0 . x. 2. (1. x2 . i = 0. See Exercise 1.. (1.1. x3 ) instead of (t. (1.17) shows that the worldline is a straight line in the spacetime. i. y.) For inertial particles this is called Newton’s ﬁrst law. Eq. Thus “straight in spacetime” includes both “straight in space” and “straight in time” (constant speed). ¨ For inertial particles we may take the parameter to be s.17). and ﬁnd that worldlines of inertial particles and light can be parameterized xi (p) = ai p + bi . Exercise 1. 1. Our third postulate for special relativity says that inertial particles and light pulses move in a straight line at constant speed in an inertial frame. (1.e. their equations of motion are xi = ai x0 + bi . a0 = 1.10.16). 2. 3. (1. i = 1.

Chapter 2

Curved Spacetimes

2.1 History of Theories of Gravity

Recall the analogy from Chapter 1: A curved spacetime is to a ﬂat spacetime as a curved surface is to a ﬂat surface. We explored the analogy between a ﬂat surface and a ﬂat spacetime in Chapter 1. In this chapter we generalize from ﬂat surfaces and ﬂat spacetimes (spacetimes without signiﬁcant gravity) to curved surfaces and curved spacetimes (spacetimes with signiﬁcant gravity). General relativity interprets gravity as a curvature of spacetime. Before embarking on a study of gravity in general relativity let us brieﬂy review the history of theories of gravity. These theories played a major role in the rise of science. Theories of gravity have their roots in attempts of the ancients to predict the motion of the planets. The position of a planet with respect to the stars changes from night to night. Some exhibit a “loop” motion, as Fig. 2.1: The position of a planet in Fig. 2.1. The word “planet” is from the ( ◦ ) with respect to the stars changes nightly. Greek “planetai”: wanderers. In the second century, Claudius Ptolemy devised a scheme to explain these motions. Ptolemy placed the Earth at the center of the universe with a planet moving on a small circle, called an epicycle, while the center of the epicycle moves uniformly on a larger circle, the deferent, around the Earth. See Fig. 2.2. By appropriately choosing the radii of the epicycle and deferent, as well as the speeds involved, Ptolemy was able to reproduce, with fair accuracy, the motions of the planets. However, this did not work exactly and so Ptolemy added epicycles to the epicycles. To explain the motions of the Moon and the known planets, Ptolemy needed 77 epicycles! This remarkable but cumbersome theory was accepted for over 1000 years.

28

2.1 History of Theories of Gravity In 1543 Nicholas Copernicus published a theory which was to revolutionize science and our perception of our place in the universe: he placed the Sun near the center of our planetary system. Copernicus retained the system of uniform circular motion using epicycles and deferents, but by placing the Sun near the center he was able to reduce greatly their number. Fig. 2.2: Ptolemy’s theory of planeIn 1609 Johannes Kepler published a tary motion. theory which discarded the whole system of epicycles and deferents and replaced it with a simpler description: the path of a planet is an ellipse with the Sun at one focus. At about the same time Galileo Galilei was investigating the acceleration of objects near the Earth’s surface. He found two things of interest for us: the acceleration is constant in time and independent of the mass and composition of the falling object. In 1687 Isaac Newton published a theory of gravity which explained Kepler’s astronomical and Galileo’s terrestrial ﬁndings as manifestations of the same phenomenon: gravity. To understand how orbital motion is related to falling motion, refer to Fig. 2.3. The curves A, B, C are the paths of objects leaving the top of a tower with greater and greater horizontal velocities. They hit the ground farther and farther from the bottom of the tower until C when the object goes into orbit! Mathematically, Newton’s theory says that a planet Fig. 2.3: Falling and in the Sun’s gravity or an apple in the Earth’s gravity orbital motion are the same. is pulled instantaneously by the central body (do not ask how!), causing an acceleration a=− κM , r2 (2.1)

where κ is the Newtonian gravitational constant, M is the mass of the central body, and r is the distance to the center of the central body. Eq. (2.1) implies that the planets orbit the Sun in ellipses, in accord with Kepler’s ﬁndings. See Appendix 8. By taking the distance r to the Earth’s center to be sensibly constant near the Earth’s surface, we see that Eq. (2.1) is also in accord with Galileo’s ﬁndings: a is constant in time and is independent of the mass and composition of the falling object. Newton’s theory has enjoyed enormous success. A spectacular example occurred in the nineteenth century. Observations of the position of the planet Uranus disagreed with the predictions of Newton’s theory of gravity, even after taking into account the gravitational eﬀects of the other known planets. The discrepancy was about 4 arcminutes – 1/8th of the angular diameter of the 29

2.1 History of Theories of Gravity moon. In 1846, U. LeVerrier, a French mathematician, calculated that a new planet, beyond Uranus, could account for the discrepancy. He wrote J. Galle, an astronomer at the Berlin observatory, telling him where the new planet should be – and Neptune was discovered! It was within 1 arcdegree of LeVerrier’s prediction. Even today, calculations of spacecraft trajectories are made using Newton’s theory. The incredible accuracy of his theory will be examined further in Sec. 3.3. Nevertheless, Einstein rejected Newton’s theory because it is based on prerelativity ideas about time and space which, as we have seen, are not correct. For example, the acceleration in Eq. (2.1) is instantaneous with respect to a universal time.

30

2.2 The Key to General Relativity A curved surface is diﬀerent from a ﬂat surface. as in special relativity. restricting them to small spacetime regions. The Newtonian theory predicts this: according to Eq. On an apple. An inertial object in gravity is in free fall. Inertial objects in the cabin move in a straight line at constant speed. This is familiar: a small region of a (perfectly) spherical Earth appears ﬂat. a simple observation of Einstein provides the key to the construction of general relativity: a small region of a curved spacetime is very much like a small region of a ﬂat spacetime. The metric postulate asserts a universal light speed and a slowing of moving clocks in local inertial frames. On the other hand. We ﬁrst discuss experimental evidence for the three postulates.1 an inertial object and the cabin accelerate the same with respect to the Earth and so they do not accelerate with respect to each other. And an accelerometer carried by the astronauts will register zero. We see this vividly in motion pictures of astronauts in orbit. objects suspended at rest remain at rest. which feels very much like an inertial force. 2. In the next three sections we shall formalize Einstein’s observation by taking our three postulates for ﬂat spacetimes. call an object inertial if it experiences no inertial forces. This is why it took so long to discover that it is not ﬂat. Accelerometers respond to the gravity. In the next three sections we shall formalize Gauss’ observation by taking the three postulates for ﬂat surfaces from Chapter 1. and then using them as postulates for curved surfaces. Forces other than gravity do not act on it. The local inertial frame postulate asserts the existence of small inertial cubical lattices with synchronized clocks to serve as coordinate systems in small regions of a curved spacetime. a simple observation by the nineteenth century mathematician Karl Friedrich Gauss provides the key to the construction of the theory of surfaces: a small region of a curved surface is very much like a small region of a ﬂat surface. However. We now rephrase Einstein’s observation: as viewed by inertial observers. No gravity is apparent in their cabin.2 The Key to General Relativity 2. Passengers in an airplane at rest on the ground or ﬂying in a straight line at constant speed feel gravity. However. To understand this. restricting them to small regions. astronauts in orbit or falling radially toward the Earth feel no inertial forces. We shall see that a curved spacetime is diﬀerent from a ﬂat spacetime. a small region of a curved spacetime is very much like a small region of a ﬂat spacetime. just as in a ﬂat spacetime. We shall include gravity as an inertial force and. a smaller region must be chosen before it appears ﬂat. The geodesic postulate asserts that inertial particles and light move in a straight line at constant speed in local inertial frames. and then using them as our ﬁrst three (of four) postulates for curved spacetimes. even though they are not moving in a straight line at constant speed with respect to the Earth. we must extend the concept of an inertial object to curved spacetimes. 31 .

Changes in this distance can be measured within 2 cm (!) by timing the return of a laser pulse sent from Earth to mirrors on the Moon left by astronauts. The weights A and B. The apparatus had a resonant period of oscillation of 24 hours so that oscillations could build up.) We may reformulate this in the language of spacetimes: the worldline of an inertial object in a curved spacetime is independent of its mass and composition. verify to extraordinary accuracy Galileo’s ﬁnding incorporated into Newton’s theory: the acceleration of a free falling object in gravity is independent of its mass and composition. If this were not so. one can wonder whether this diﬀerence between the Earth 32 .4.4: Masses A and B accelerate the Earth’s rotation. This energy is called the Earth’s gravitational binding energy . supported by a quartz ﬁber. The measurements show that the relative acceleration of the Earth and Moon is no more than a part in 1013 of their mutual acceleration toward the Sun. Thus the separated pieces have more total mass than the Earth. in the opposite direction twelve hours later. 2. (2. this energy is equivalent to mass. Since the gravitational binding energy mass is in some sense diﬀerent from “ordinary” mass.1) does not change by more than 1 part in 1012 per year. 2.1.6 parts in 1010 . Due to Fig. with the Earth. By Einstein’s principle of the equivalence of mass and energy (E = mc2 ). In Braginsky’s experiment the diﬀerence in the acceleration of the weights toward the Sun was no more than one part in 1012 of their mutual acceleration toward the Sun. Dicke and of V. Braginsky. Imagine disassembling to Earth into small pieces and separating the pieces far apart. A related experiment shows that the Earth and the Moon accelerate the same in the Sun’s gravity. performed in the 1960’s. The experiment also shows that the Newtonian gravitational constant κ in Eq. This is part of the lunar laser experiment. The constant also appears in Einstein’s ﬁeld equation. We can understand the principle of their experiments from the simpliﬁed diagram in Fig. 2. A planned satellite test (STEP) will test the equality of accelerations to one part in 1018 . There is another diﬀerence between the Earth and the Moon which might cause a diﬀerence in their acceleration toward the Sun. Dicke and Braginsky used the Sun’s gravity. (See Sec. The separation requires energy input to counteract the gravitational attraction of the pieces. B. (2. The diﬀerence is small – 4. Dicke and Braginsky used various substances with various properties for the weights.2. despite the huge diﬀerence in their masses.2 The Key to General Relativity Experiments of R.27). The geodesic postulate will incorporate this by not referring to the mass or composition of the inertial objects whose worldlines it describes. Eq. It is 23 times smaller for the Moon. Any difference in their acceleration toward the Sun would cause a twisting of the ﬁber. are. the twisting would be the same toward the Sun. then there would be unexpected changes in the Earth-Moon distance. in free fall around the Sun.

(It is desirable to test this directly by performing the experiment in orbit. Suppose the tower is momentarily at rest when gamma radiation is emitted. Loosely speaking.1. necessary for clock synchronization.3 that the assumption Eq. Rebka in 1960 and then more accurately by Pound and J. Snider in 1964.6). This is something that the Dicke and Braginsky experiments cannot test. The last experiment we shall consider as evidence for the three postulates is the terrestrial redshift experiment. (We ignore the small distance the tower moves during the ﬂight of the radiation. These eﬀects are far too small to be detected by the experiment. we may say that since light behaves “properly” in a small inertial lattice.2. This assumption fails for clocks at the top and bottom of the tower. c c (2. the tower has (upward) acceleration g. We will show shortly that the same redshift would be observed with a tower having acceleration g in an inertial frame in a ﬂat spacetime. is equivalent to a zero redshift between the clocks. in the inertial frame. the acceleration of Earth’s gravity.4). We now calculate the Doppler redshift for a tower with acceleration g in an inertial frame. This is another example of small regions of ﬂat and curved spacetimes being alike. in a small inertial lattice falling radially toward Earth. We showed in Sec. an observer at the top of the tower would see a clock at the bottom run slowly. within an experimental error of about 1%.4). We now show that the experiment provides evidence that clocks at rest in a small inertial lattice can be synchronized. 1. This is a gravitational redshift. The radiation travels a distance h. L. is valid for clocks at rest in a small inertial lattice. (1. See Exercise 2. V. the height of the tower. A. (1. just as with a tower at rest in an inertial frame. 33 .) In this time the tower acquires a speed v = gt = gh/c in the inertial frame. Clocks at rest at diﬀerent heights in the Earth’s gravity run at diﬀerent rates! Part of the result of the Hafele-Keating experiment is due to this. light accelerates the same as matter in gravity. Radiation received at the top of the tower was redshifted: z = 2.2 The Key to General Relativity and Moon causes a diﬀerence in their acceleration toward the Sun. Thus clocks at rest in a small coordinate lattice on the ground cannot be (exactly) synchronized. The lunar laser experiment shows that this does not happen. Substituting numerical values in Eq.) In this way. We shall also ignore the time dilation of clocks in the moving tower and the length contraction – see Appendix 4 – of the tower. this speed causes a Doppler redshift z= gh v = 2. According to the discussion following Eq.5 × 10−15 .) Thus the radiation takes time t = h/c to reach the top of the tower. It was ﬁrst performed by R. Thus it is reasonable to assume that there would be no redshift with a tower at rest in a small inertial lattice in gravity. necessary for synchronizing clocks at rest in the coordinate lattice of an inertial frame. (1. h = 2250 cm. (For clarity we do not take c = 1. The experimenters put a source of gamma radiation at the bottom of a tower. From Exercise 1. Pound and G.6. the experiment provides evidence that the condition Eq.2) In the experiment. In the experiment.

9. Using this system. Exercise 3. the gravitational redshift for towers accelerating in inertial frames is the same as the Doppler redshift for towers accelerating in small inertial lattices near Earth. Exercise 3. Show that the diﬀerence between the clocks due to the gravitational redshift is ∆sa − ∆sg = gh∆t. 34 .6×10−7 sec.2 gives the value of z measured in the terrestrial redshift experiment.0 × 10−7 sec.9 gives a diﬀerence of 3. the eﬀects are 10. The Global Positioning System (GPS) is a system of 24 Earth satellites. In fact. Substitute values to obtain ∆sa −∆sg = 1.3 shows that a rigorous calculation in general relativity gives the same result.1 to function properly. (2.1.2. Exercise 2.2). The system must take into account the special and general relativistic eﬀects of Exercise 2. Adding this to the time dilation diﬀerence of Exercise 1. Suppose that h = 10 km.000 times too large to be ignored. a user on Earth can determine their position to within a few meters with a hand held device.4 shows that a rigorous calculation in general relativity also gives Eq.2 The Key to General Relativity 2. Let h be the height at which the airplane ﬂies in the simpliﬁed Hafele-Keating experiment of Exercise 1.

they merely serve to label the points of the surface. In general. The coordinates are assigned by the usual parameterization x = R sin φ cos θ. There are. deﬁned over the entire surface.5: A local planar frame. Surface dwellers call a small (nearly) square coordinate grid a local planar frame at P . y 2 ) varies. θ) on a sphere of radius R. θ) nates (y . the coordinates will have no geometric meaning. in general. In smaller regions around P . z) varies on the surface.2. y = y(y 1. y 2 ) . y 2 ). y 2 ). no natural global coordinate systems to single out in a curved surface as planar frames are singled out in a ﬂat surface. 2. then at ﬁrst they will ﬁt together well. But. The only restrictions are that diﬀerent points on a sphere. (x. 2. (2.3) As (y 1. y 2 ) to the point (x. Assign coordinates (y 1. they need global coordinates. One common way for us (but not surface dwellers) to attach global coordinates to a curved surface is to parameterize it in three dimensional space: x = x(y 1. owing to the curvature of the surface. z) on the surface given by Eq. For example. However. in order to study the surface as a whole. See Fig.3 The Local Inertial Frame Postulate Suppose that curved surface dwellers attempt to construct a square coordinate grid using rigid rods constrained (of course) to their surface.4) 35 . y 2 ) = (φ. y. y ) in an arbitrary manner. the grid must become more square. (2. Thus they attach global coordi1 2 Fig. y. as the grid gets larger the rods must be forced a bit to connect them. Fig.6 shows spherical coordinates (y 1.5.3 The Local Inertial Frame Postulate 2. We shall see that local planar frames at P provide surface dwellers with an intuitive description of properties of a curved surface at P . This will cause stresses in the lattice and it will not be quite square. The local planar frame postulate for a curved surface A local planar frame can be constructed at any point P of a curved surface with any orientation. y = R sin φ sin θ. If the rods are short enough. must have diﬀerent coordinates and nearby points must receive nearby coordinates. z = z(y 1. Fig. z = R cos φ. 2. 2. (2.6: Spherical coordinates (φ.3). where P is the point at the origin of the grid.

2. This will cause stresses in the lattice and it will not be quite cubical. Local inertial frames are in free fall. 2. assume for simplicity that their cabin is falling radially toward Earth. 1. then at ﬁrst they will ﬁt together well.10. A small (nearly) cubical lattice with (nearly) synchronized clocks is called a local inertial frame at E.) In the last section we saw that inertial objects in an astronaut’s cabin behave as if no gravity were present. they will not behave ex:actly as if no gravity were present.3 The Local Inertial Frame Postulate Surface dwellers on the sphere could assign these coordinates without leaving the surface. an object initially at rest near the top of the cabin will slowly separate from one initially at rest near the bottom. Suppose that astronauts in a curved spacetime attempt to construct an inertial cubical lattice using rigid rods. Actually. See Fig. See Fig. (Technically. Actually. 36 . 2. (We shall see in Sec. two objects initially at rest at the same height will slowly move toward each other as they both fall toward the center of the Earth.7. To see this.2. The statement given will suﬃce for us. For example. However. the circle of latitude at distance r from the North Pole is assigned the coordinate φ = r/R. due to small Fig.e. y 2 ). we can hope that the procedure will work with as small an error as desired by restricting the lattice to a small enough region of a spacetime. i. These changes in velocity are called tidal accelerations. and the rods cannot all be inertial. If the rods are short enough. in smaller regions of spacetime. Thus. free falling cabin. In smaller regions around E. the postulate should state that a curved surface is a two dimensional manifold . the lattice is more cubical and the clocks are more nearly synchronized.3 will not quite work. (Why?) They are caused by small diﬀerences in the Earth’s gravity at diﬀerent places in the cabin. Inertial objects in the cabin do not accelerate exactly the same with respect to the Earth because they are at slightly diﬀerent distances and directions from the Earth’s center. 2. the lattice will have to resist tidal accelerations.7: Tidal accelerations in a radially diﬀerences in the gravity at diﬀerent places in the lattice. But as the grid gets larger.) The global coordinate postulate for a curved surface The points of a curved surface can be labeled with coordinates (y 1 .6 that surface dwellers can determine great circles and R without leaving the surface.. an attempt to synchronize the clocks with the one at the origin with the procedure of Sec. In addition. where 2πR is the distance around a great circle. They become smaller in smaller regions of space and time. where E is the event at the origin of the lattice when the clock there reads zero. In the last section we saw that the terrestrial redshift experiment provides evidence that clocks in a small inertial lattice can be synchronized.

in order to study a curved spacetime as a whole. G.3 The Local Inertial Frame Postulate The local inertial frame postulate for a curved spacetime A local inertial frame can be constructed at any event E of a curved spacetime. x) with (t + x. no natural global coordinates to single out in a curved spacetime. We can then put the metric and geodesic postulates of special relativity in the same global form that we shall obtain for these postulates for curved spacetimes.2. The only restrictions are that diﬀerent events must receive diﬀerent coordinates and nearby events must receive nearby coordinates. with any orientation. we need global coordinates. Thus the mathematics necessary to describe curved spacetimes intrinsically was waiting for Einstein when he needed it. 37 . they merely serve to label the events of the spacetime. His work was extended by several mathematicians. in general. as inertial frames were singled out in a ﬂat spacetime. t − x). This global form of the postulates is unintuitive and complicated but is necessary to carry out calculations in the theory. In general. Often one of the coordinates is a “time” coordinate and the other three are “space” coordinates. and with any inertial object at E instantaneously at rest in it. We shall ﬁnd that local inertial frames at E provide an intuitive description of properties of a curved spacetime at E. in a ﬂat spacetime it is sometimes useful to replace the coordinates (t.e. Thus we attach global coordinates in an arbitrary manner. In the next two sections we give the metric and geodesic postulates of general relativity. without describing it as curved in a higher dimensional ﬂat space. The global coordinate postulate for a curved spacetime The events of a curved spacetime can be labeled with coordinates (y 0 . deﬁned over the entire spacetime. It is remarkable that we shall be able to describe curved spacetimes intrinsically. We ﬁrst express the postulates in local inertial frames. For example. y 1 . but this is not necessary. Riemann generalized Gauss’ mathematics to curved spaces of higher dimension in 1854. the coordinates will have no physical meaning. There are.. B. We then translate the postulates to global coordinates. We do not usually use arbitrary coordinates in ﬂat spacetimes because inertial frames are so much easier to use. However. This local form of the postulates gives them the same physical meaning as in special relativity. Gauss created the mathematics necessary to describe curved surfaces intrinsically in 1827. We can use arbitrary global coordinates in ﬂat as well as curved spacetimes. y 2 . We do not have this luxury in curved spacetimes. y 3 ). i.

Thus Eq.8) = gjk (y) dy j dy k . (2. dx2 ) in a local planar frame at P .5) in terms of global coordinates. consider a function f (y 1 . We may think of the (xi ) coordinates as functions of the (y j ) coordinates. 38 . Let Q have coordinates (dx1 .6) Henceforth we use the Einstein summation convention by which an index which appears twice in a term is summed without using a Σ. (2. Then ds2 = (dx1 )2 + (dx2 )2 .5) holds only for inﬁnitesimal distances from P . k) ∂y k dy j dy k (2.4 The Metric Postulate 2. local form Let point Q have coordinates (dx1 .4 The Metric Postulate Fig. Then we may write the diﬀerential df = (∂f /∂y i ) dy i . (2.7) As another example of the summation convention. j.5 shows that local planar frames provide curved surface dwellers with a convenient way to express inﬁnitesimal distances on a curved surface. (2. (2.5) can be written 2 ds = m. dx2 ) in a local planar frame at P . We now express Eq. (2. y 2 + dy 2 ) in a global coordinate system. Let P and Q be neighboring points on a curved surface with coordinates (y 1 . The metric postulate for a curved surface. y 2 ). From Eq. (2. Then Eq.7). y = r sin θ. y 2 ) and (y 1 + dy 1 . n.n=1 2 ◦ fmn dxm dxn . This gives meaning to the partial derivatives ∂xi/∂y j . (2. Set the matrix ◦ f ◦ = (fmn ) = 1 0 0 1 .6) becomes ◦ ds2 = fmn dxm dxn . 2. just as cartesian coordinates in the plane are functions of polar coordinates: x = r cos θ.5) Even though the local planar frame extends a ﬁnite distance from P .2. the distance from P to Q is ◦ ds2 = fmn dxm dxn ◦ = fmn = ∂xm j dy ∂y j m n ◦ ∂x ∂x fmn j ∂y ∂y k ∂xn k dy (sum on m. Eq. Let ds be the distance between the points.

Show that for a local planar frame whose x1 -axis coincides with a circle of latitude.2.4) to convert ds2 = dx2 + dy 2 + dz 2 to (φ. but as a single object v which represents a magnitude and direction (an arrow). 1 Exercise 2. We should not think of a vector as its components (vi ) . In a given coordinate system the metric acquires components. dx1 = R sinφ dθ and dx2 = Rdφ.10) (gjk (y i )) is called metric of the surface with respect to the coordinates (y i ). (2. with metrics (gjk (y i )) and (¯pq (¯i )). (2. Then there is a symmetric matrix (gjk (y i )) such that ds2 = gjk (y i ) dy j dy k . Consider the hemisphere z = (R2 − x2 − y 2 ) 2 . z) on the hemisphere. global form Let (y i ) be global coordinates on the surface. y.4) is R2 0 . y) to the point (x. ∂ yp ∂ yq ¯ ¯ (2. Let (y i ) and (¯i ) be two coordinate systems on the same y surface. In a given coordinate system the vector acquires components. Assign coordinates (x. (2. to global coordinates: The metric postulate for a curved surface. (2. Use a local planar frame at each point of the surface in this manner to translate the local form of the metric postulate.4 The Metric Postulate where we have set ◦ gjk (y) = fmn ∂xm ∂xn . we should not think of the metric as its components (gjk ) .2. (2. The components will be diﬀerent in diﬀerent coordinate systems. ∂y j ∂y k . Show that g y gpq = gjk ¯ Hint: See Eq. θ) coordinates on the sphere in Eq. b.12) 39 . Exercise 2.3. θ) coordinates. Similarly.11) 2 0 R sin2 φ Do this in two ways: a.4.10. Express your answer as a matrix. Use Eq. 2.9) ◦ Since (fmn ) is symmetric.5). See Fig. The components will be diﬀerent in diﬀerent coordinate systems. Show that the metric for the (φ.8). (2. so is (gjk ). ∂y j ∂y k (2. Eq. By converting from the metric of a local planar frame. but as a single object g which represents inﬁnitesimal distances. Exercise 2. Find the metric in this coordinate system. Hint: Use z 2 = R2 −x2 −y 2 to compute dz 2 . Let ds be the distance between neighboring points (y i ) and (y i + dy i ).

Then there is a symmetric matrix (gjk (y i )) such that ds2 = gjk (y i ) dy j dy k . local form Let event F have coordinates (dxi ) in a local inertial frame at E.13) can be written ◦ ds2 = fmn dxm dxn .10). (2.14) where 1 0 0 0 0 −1 0 0 ◦ . If neighboring events (y i ) and (y i + dy i ) are lightlike separated.) This is another instance of the key to general relativity: a small region of a curved spacetime is very much like a small region of a ﬂat spacetime. let ds = 0. The metric postulate for a curved spacetime. f ◦ = (fmn ) = 0 0 −1 0 0 0 0 −1 Using Eq. Eq. If the events are timelike separated. global form Let (y i ) be global coordinates on the spacetime. let ds be the time between them as measured by a clock moving between them.14). (2.15) 40 . the calculation Eq. (2. (2. Then ds2 = (dx0 )2 − (dx1 )2 − (dx2 )2 − (dx3 )2 . which produced the global form of the metric postulate for curved surfaces. now produces the global form of the metric postulate for curved spacetimes. If the events are timelike separated. (See the discussion following Eq. (2. (1. If E and F are lightlike separated. We now translate the metric postulate to global coordinates.4 The Metric Postulate The metric postulate for a curved spacetime. let ds = 0. let ds be the time between them as measured by any (inertial or noninertial) clock.2. (2.13) The metric postulate asserts a universal light speed and a formula for proper time in local inertial frames.8). (gjk (y i )) is called the metric of the spacetime with respect to (y i ).

5. Fig. To traverse a geodesic. Then in every local planar frame at P xi (P ) = 0. a surface dweller need only always walk “straight ahead”. We ﬁrst need to know that the metric g = (gij ) has an inverse g−1 = (g jk ).5 The Geodesic Postulate Curved surface dwellers ﬁnd that some curves in their surface are straight in local planar frames. The Γi . i = 1. 2. θφ φθ θθ The remaining Christoﬀel symbols are zero.16) to global coordinates y to obtain the global form of the geodesic equations. where t means “transpose”. Let the matrix a = (∂xn /∂y k ). Exercise 2.6. (2. 2. (2. (2. ¨ (2. Show that the inverse matrix a−1 = (∂y k /∂xj ). They call these curves geodesics. (2. Exercise 2. As with the geodesic postulate for ﬂat surfaces Eq.18) 41 . global form Parameterize a geodesic with areclength s .8: The equator is the only circle of latitude which is a geodesic. (2. 2. Γi = jk 1 2 g im [∂k gjm + ∂j gmk − ∂m gjk ] .8 shows that the equator is a geodesic but the other circles of latitude are not. Then in every global coordinate system y i + Γi y j y k = 0. You should not try to assign a geometric meaning to the Christoﬀel symbols. are jk kj jk functions of the coordinates.15).16) We now translate Eq.17) Note that Γi = Γi . ¨ jk ˙ ˙ i = 1. c. (1. Introduce the notation ∂k gim = ∂gim /∂y k .5 The Geodesic Postulate 2. they should simply be thought of as what appears when the geodesic postulate is translated from its local form Eq. given that it is constrained to the surface. Let point P be on the worldline.11) Γφ = − sinφ cos φ and Γθ = Γθ = cot φ.9) can be written g = at f ◦ a. Show that for the metric Eq. local form Parameterize a geodesic with areclength s . we have The geodesic postulate for a curved surface. like the gjk .2. (2. a. Show that Eq. Deﬁne the Christoﬀel symbols: Fig. A geodesic is as straight as possible. 2. Prove that g−1 = a−1 (f ◦ )−1 (a−1 )t .16) (which has an evident geometric interpretation) to its global form (which does not): The geodesic postulate for a curved surface. b.

local form Worldlines of inertial particles and pulses of light can be parameterized so that if E is on the worldline. Of course. (1. (2. called geodesic coordinates. all great circles on a sphere are geodesics. Exercise 2. Exercise 2. 2. 1.7. The geodesic postulate asserts that inertial particles and light move in a straight line at constant speed in a local inertial frame.2. surface dwellers assume that the metric of a local planar frame at P satisﬁes Eq.9. Use the result of Exercise 2.18).19). (fmn (P )) = f ◦ .5 The Geodesic Postulate Appendix 9 translates the local form of the postulate to the global form. Since a local planar frame at P is constructed to approximate a planar frame as closely as possible near P . 2. Appendix 10 shows that there are coordinates. A function with a zero derivative at a point is not changing much at the point. y) = (0. (2.19) states that f stays close to f ◦ near P .17)). satisfying this relationship and also ∂i fmn (P ) = 0 (2. Show that Eq. i. Show that the equator is the only circle of latitude which is a geodesic. Show that the metric of Exercise 2.9: The path of an inertial particle in a time is as straight as possible (in lattice stuck to the Earth and in an inertial space and time – see the remarks lattice.19) at (x. but looks curved when viewed in an “inappropriate” coordinate system.) The worldline of an inertial particle or pulse of light in a curved space.Fig.7). (2. Einstein’s “straightest worldline in a curved spacetime” description is very diﬀerent from Newton’s “curved path in a ﬂat space” description. n. ¨ For inertial particles we may take the parameter to be s. (2.18) reduces to Eq.9. In this sense Eq. following Eq. Before using the geodesic equations you must parameterize the circles with s. Let f = (fmn (x)) represent the metric in this coordinate system. (1. The dots are at equal time intervals. See Fig. i = 0. The worldlines are called geodesics. (2. The geodesic postulate for a curved spacetime.19) for all m. given that it is constrained to the spacetime.3 satisﬁes Eq. The geodesic is straight in local inertial frames. Exercise 2. 42 (2. A local planar frame at P extends to a ﬁnite region around P .6. According to Eq.16) for local planar frames. (See the remarks following Eq.20) .8. The translation requires an assumption. 3. (2. then in every local inertial frame at E xi (E) = 0. 2. 0) .

19) for the same reasons as given above for local planar frames. ¨ jk ˙ ˙ i = 0.2. Then Appendix 9 translates the local form of the geodesic postulate for curved spacetimes to the global form: The geodesic postulate for a curved spacetime. 3. 43 .21) For inertial particles we may take the parameter to be s. global form Worldlines of inertial particles and pulses of light can be parameterized so that in every global coordinate system y i + Γi y j y k = 0. 1. (2. 2.5 The Geodesic Postulate We now assume that the metric of a local inertial frame at E satisﬁes Eq. (2.

Thus C(r) > 2πr and so K ≤ 0.17). 2. in general K varies from point to point in a curved surface. From Fig. Connect all the points at distance r from P along the geodesics. A quick calculation shows that K = 1/R2 . Take the Earth to be perfectly spherical. ness of solutions of systems of diﬀerential equations. Let C(r) be the circumference of the circle.6 The Field Equation Previous sections of this chapter explored the similarities between small regions of ﬂat and curved surfaces and between small regions of ﬂat and curved spacetimes. 2. . But distances determine geodesics: disa pseudosphere. Despite the examples. This section explores the diﬀerences. 2 Hint: Use a basic theorem on the existence and unique. Show that a map of a region of the Earth must distort distances.10: K = 1/R on a sphere. . pass geodesics through a point P in every direction.18).11. The local forms of our curved surface postulates show that in many ways a small region of a curved surface is like a small region of a ﬂat surface. . 2. Exercise 2. Thus we learn an important fact: distances determine K.10. Show that there is a unique geodesic through every point of a curved surface in every direction. The curvature is a diﬀerence between regions of a sphere and a ﬂat surface which does not vanish as the regions become smaller.10. (2.11 is a pseudosphere. The surface of revolution of Fig.Fig. r is the radius of a circle. Make no calculations.2. K = 0 for a ﬂat surface.14 shows that K = −1/R2 . which determines the Christoﬀel symbols Eq. C(r) < 2πr for a sphere and so K ≥ 0. (2.11: K = −1/R on nition of K. C.6 The Field Equation 2. which determine the geodesics Eq. where R is a constant. Surface dwellers might suppose that all diﬀerences between the regions vanish as the regions become smaller.22) Clearly. 44 .10. tances determine the metric. 2. From Fig. Thus K is measurable by surface dwellers. For surface dwellers. Exercise 2. This is not so. Distances and geodesics are involved in the deﬁ2 Fig. Exercise 2. Deﬁne the curvature K of the surface at P : K= 3 2π r − C(r) lim . we ﬁnd C(r) = 2πR sin φ = 2πR sin(r/R) = 2πR r/R − (r/R)3 /6 + . The horizontal “circles of latitude” are concave inward and the vertical “lines of longitude” are concave outward. To see this. π r→ 0 r3 (2. 2.

6 The Field Equation We can roll a ﬂat piece of paper into a cylinder without distorting distances along curves in the paper and thus without changing K. how it is “curved”. In a ﬂat spacetime. (2. and inertial particles in a curved (or ﬂat) spacetime. 45 . To see this.14. This is not so. Our ﬁnal postulate for general relativity. then the surface of revolution in Exercise 2. Thus surface dwellers could not detect the curvature seen from the outside. 2.1). The local forms of our curved spacetime postulates show that in many respects a small region of a curved spacetime is like a small region of a ﬂat spacetime.11. We might suppose that all diﬀerences between the regions vanish as the regions become smaller.23). the cylinder is curved. If y = 1 and z = u. (2. the equation determines the “shape” of a spacetime. In a ﬂat spacetime ∆ a = 0 for inertial particles. (2. Show that K = 0 for a cylinder using Eq. ∆ a/∆ r → 0. viewed from within (and remember. Use Eq. Use ds2 = dr2 + r2 dθ2 + dz 2 to show that the metric is f 2 +h2 0 0 f2 u . we must know the metric of the spacetime. The metric and geodesic postulates describe the behavior of clocks. Loosely speaking.13. If g12 = 0. the rolling does not distort distances in the paper. Let ∆ r be the small distance between objects at the top and bottom of the cabin and let ∆ a be the small tidal acceleration between them. 1 Exercise 2. the ﬁeld equation.7. z) be cylindrical coordinates and parameterize the surface with coordinates (u. Exercise 2. then the surface of revolution in Exercise 2.12.2. However. Let (r. Here is a diﬀerence between regions of the spacetime in the cabin and a ﬂat spacetime which does not vanish as the regions become smaller. If y = Re−u and z = R 0 (1 − e−2t ) 2 dt.11). However. and so K = 0 seems “wrong”. we are describing curved surfaces and spacetimes without reference to a higher dimensional space). as ∆ r → 0 : from / Eq. The formula expressing K in terms of distances was given by Gauss. 2. but ∆ a → 0 as ∆ r → 0 . Thus K = 0 for the cylinder. Show that Eq. (2. Show that K = −1/R2 for the pseudosphere.23) gives K = 1/R2 for a sphere of radius R.13 is a cylinder.23) Exercise 2. determines the metric. da/dr = 0. θ. da/dr = 2κM/r3 . In the cabin ∆ a = 0 . refer to Fig. Exercise 2. Thus K “should” be zero for the cylinder.15.13 is the pseudosphere of Fig. light. z = h(u) about the z-axis. Generate a surface of revolution by rotating the parameterized curve y = f (u). But to apply these postulates. θ). then K = − (g11 g22 ) 1 −2 2 2 ∂1 g112 ∂1 g22 + ∂2 g222 ∂2 g11 −1 1 −1 1 ∂i ≡ ∂/∂y i . (2. Viewed from the outside.

existing independently of any coordinate system. and also pressures. Gas in interstellar space which is thin enough so that particle collisions are infrequent is dust. called dust. But for our purposes it will be suﬃcient to consider only matter of a special form. contribute to T. jk As with the metric g . The Ricci tensor is entirely determined by the metric. All forms of matter and energy.24): the quantity determined by the metric is the Einstein tensor G = R − 1 R g.25) Don’t panic over this convoluted deﬁnition: You need not have a physical understanding of the Ricci tensor. It represents the source of the gravitational ﬁeld in general relativity. Deﬁne T jk = ρ dy j dy k .6 The Field Equation We constructed the metric in Sec.26) We can now state our ﬁnal postulate for general relativity: 46 . (2. including electromagnetic ﬁelds. Thus. We shall see that it contains information about the curvature K of two dimensional surfaces in four dimensional spacetime. 2. Schematically it reads quantity determined by metric = quantity determined by mass/energy . 2 The right side of the ﬁeld equation is given by the energy-momentum tensor T. ds ds (2. there is a relationship between the metric of a spacetime and the distribution of mass in the spacetime. There is obviously a relationship between the motion of local inertial frames and the distribution of mass in a curved spacetime. Let ds be the time between the events measured by a clock moving with the element. We can now specify the left side of the schematic ﬁeld equation Eq. it involves second derivatives of the gjk (because the Γi involve the ﬁrst derivatives). Deﬁne the curvature scalar R = g jk Rjk . If there is an element of dust at E. each inﬁnitesimal element of matter is inertial and thus moves on a geodesic.24) To specify the two sides of this equation. but which in a given coordinate system acquires components Rjk .4 using local inertial frames. computers can readily calculate them for us.2. (2. we need several deﬁnitions. And while the Rjk are extremely tedious to calculate by hand. we will use R to designate the Ricci tensor as a single object. let (y i + dy i ) be a neighboring event on its worldline. We now deﬁne T for dust. Like K. The ﬁeld equation gives this relationship. tp jk jp tk jp jk (2. In dust. Let ρ be the density of matter at E as measured by an observer moving with the matter. Deﬁne the Ricci tensor Rjk = Γp Γt − Γp Γt + ∂k Γp − ∂p Γp . Consider an event E with coordinates (y i ).

4. Gjk = Rjk − 1 Rgjk .6 The Field Equation The ﬁeld equation G = −8πκT.2. The equation is a system of second order nonlinear partial diﬀerential equations with the gij as unknowns. the gjk transform covariantly.1 Subscripted components ajk transform covariantly: apq = ajk ¯ ∂y j ∂y k . and the ﬁeld equation are applicable in all coordinate systems. i. (2. represented by T. to the density of matter and energy in the spacetime. Tensor algebra and calculus are powerful tools for computations in general relativity. By convention. transforms covariantly. the curvature of spacetime. represented by G. then the ﬁeld equation is Rjk − 1 Rgjk = 0. You should be impressed with the power of this mathematics! If ρ = 0 at some event. event by event. Before we can use the ﬁeld equation we must deal with a technical matter. ∂ yp ∂ yq ¯ ¯ (2.. 2 Substitute R = 0 into the ﬁeld equation to obtain R = 0. without reference to a higher dimensional ﬂat space in which they are curved. the geodesic equations. Thus the left side of the ﬁeld equation. The Rjk also transform covariantly.1). It relates. (Do not attempt to verify this at home!) The Ricci scalar is the same in all coordinate systems.28) At events in a spacetime where there is no matter we may use this vacuum ﬁeld equation. 47 .29) According to Exercise 2. (2.e. The ﬁeld equation is the centerpiece of Einstein’s theory. Multiply 2 this by g jk : 1 1 0 = g jk Rjk − 2 R gjk = g jk Rjk − 1 R gjk g jk = R − 2 R 4 = −R . ∂y j ∂y k (2. The indices on the energy-momentum tensor are superscripts: T jk .30) 1 This is a convention of tensor algebra. (2. the placement of indices indicates how components transform under a change of coordinates.27) Here κ is the Newtonian gravitational constant of Eq. The indices on the metric and the Ricci tensor are subscripts: gjk and Rjk . The metric. 2 Superscripted components ajk transform contravariantly: apq = ajk ¯ ∂ yp ∂ yq ¯ ¯ . But we do not need them for a conceptual understanding of the theory. They are examples of our ability to describe curved spacetimes intrinsically.

i. According to the local intertial frame postulate.2. Consider ﬁrst the right side of Eq. if ρ = 0. Then we can take the components of the ﬁeld equation to be Gjk = −8πκT jk . (2. Show that if the Gjk transform covariantly. Then G00 = −(K12 + K23 + K31 ).) The 00 component of the ﬁeld equation gives a simple and elegant relationship between the curvatures K of certain surfaces through an event and the density ρ at the event. based on reasonable assumptions.24) to the ﬁeld equation Eq. (2. since the two sides transform the same between coordinate systems. but it should be convincing enough to make us anxious to confront the theory with experiment in the next chapter. If this equation is valid in any one coordinate system. It is reasonable to involve the density of matter. (1. We do not give a construction of these coordinates. Second. (2. (2. (Alternatively.14). ∂i fmn (E) = 0. Let K12 be the curvature of the surface formed by holding the time coordinate x0 and the spatial coordinate x3 ﬁxed. it need not be in another. we can lower the contravariant indices to covariant indices: Tmn = gmj gnk T jk .. from Appendix 4. we cannot take Gjk = −8πκT jk as the components of the ﬁeld equation: even if this equation were true in one coordinate system.16. According to Eq. Show that the T jk transform contravariantly.24) to be a local equation. Exercise 2. the metric of a Fermi normal coordinate system satisﬁes further ∂0 ∂i fmn (E) = 0.19). x1 and x2 . the element is instantaneously at rest in some local intertial frame at E.6 The Field Equation Exercise 2. This will by no means be a proof of the equation.17. Consider the element of matter at an event E.27). it will equate the two quantities event by event.24). First. The reader may turn to the next chapter without loss of continuity. the mass of a body moving with speed v increases by 1 a factor (1 − v 2 )− 2 . We take Eq. then the Gmn transform contravariantly. this factor is dx0 /ds. (2. then K12 + K23 + K31 = 0. The metric f of a local inertial frame at E satisﬁes Eq.12). And in a spacetime. it is true in all. The solution is to raise the covariant indices to contravariant indices: Gmn = g mj g nk Gjk . In particular. To obtain it we must use Fermi normal coordinates. while varying the other two spatial coordinates. (2. We close this chapter with a plausibility argument. and use the equivalent equation Gjk = −8πκTjk . which leads from the schematic ﬁeld equation Eq. Since the Gjk and T jk transform diﬀerently. then. In special relativity there are two eﬀects aﬀecting the density of moving matter. Thus from the ﬁeld equation K12 + K23 + K31 = 8πκρ. which we take to have Fermi normal coordinates.e. The metric of a geodesic coordinate system satisﬁes in addition Eq. (fmn (E)) = f ◦ . an inertial body contracts in its direction of motion by the same factor and does not contract in directions perpendicular to its direction of 48 .

49 . (2. Since the T jk transform contravariantly.34) becomes Λg = 0.4.31) where ρ is the density measured by an observer moving with the matter.32) This quantity represents matter in special relativity. namely. We will show at the end of Sec. Thus the ﬁeld equation is of the form Gjk = −8πκT jk .25). the right side of Eq. (2.24) is the energymomentum tensor T. A local inertial frame is in many respects like an inertial frame in special relativity. (We will revisit this assumption in Sec. Thus Eq. Thus Eq. (2.27).6 The Field Equation motion. depends only on the gjk and their ﬁrst and second derivatives.17). As emphasized in Chapter 1. 4.32). 2. like the curvature K of a surface. (iv) Assume that Einstein’s theory agrees with Newton’s when Newton’s is accurate. the coordinate diﬀerence dx0 has no physical signiﬁcance in and of itself.) In a ﬂat spacetime the Christoﬀel symbols Eq. (2. (2.24) with Eq. G depends on g. (2. (2.2. According to a mathematical theorem of Lovelock.24) must transform contravariantly. ds ds (2. Thus the density of moving matter is T 00 = ρ dx0 dx0 . ds ds (2. our assumptions already imply that there is no loss of generality in taking the ﬁeld equation to be of the form 1 A R − 2 gR + Λg = −8πκT . (2.34) Here A and Λ are constants. Assume that the left side of Eq. (For k = 0 this expresses conservation of mass and for k = 1. (iii) Assume that a spacetime without matter is ﬂat.24).31) is only one component of a whole: T jk = ρ dxj dxk . 3 it expresses conservation of the k component of momentum. (2. (i) From Eq. for weak gravity with small velocities. (2. (2.) Assume that ∂T jk /∂xj = 0 at the origin of local inertial frames. Thus at the origin of a local inertial frame we replace the right side of Eq.5 that this requires that A = 1.33) (The factor −8πκ was inserted for later convenience. (2.34) is precisely the ﬁeld equation Eq. Thus Λ = 0. We denote it Gjk in anticipation that it will turn out to be the Einstein tensor. (2.) We shall make four assumptions which will uniquely determine the Gjk . Thus Eq. Transforming to global coordinates. 4. (ii) In special relativity ∂T jk /∂xj = 0.24). and the curvature scalar R all vanish. (2. the left side of Eq. the Ricci tensor Eq.

4 solar masses or less. Sirius. If it accretes suﬃcient matter. For a star of ∼ 1. Clouds of interstellar gas and dust are a major component of our Milky Way galaxy.to about 1000 km . For example. We thus begin with a brief survey of the relevant aspects of stellar evolution. a Type Ia supernova has a luminosity over a billion times that of the Sun. At its brightest. Suppose some perturbing force causes a cloud to begin to contract by self gravitation. A star is born! The star will continue to shine for millions or billions of years until its nuclear fuel runs out and it begins to cool.and its density will thus increase by a factor of 106 .1 Stellar Evolution Several applications of general relativity described in this chapter involve observations of stars at various stages of their life. the radius of the Sun will decrease by a factor of 102 . it sometimes accretes matter from its companion. the brightest star in the sky. The subsequent evolution of the star depends on its mass. called a Type Ia supernova. If the white dwarf is a member of a binary star system. it will erupt in a runaway thermonuclear explosion. destroying the star. Then the contraction will begin again. For example. is part of a double star system.but not until enormous densities are reached. They are common. For a larger cloud. The heat from these reactions will increase the pressure in the cloud and stop the contraction. then gas pressures and/or mechanical forces will stop the contraction and a planet sized object will result. If the mass of the cloud is small enough. Its dim companion is a white dwarf. 50 .Chapter 3 Spherically Symmetric Spacetimes 3. the contraction will be stopped by a phenomenon known as degenerate electron pressure . The cloud will continue to contract and become hotter until thermonuclear reactions begin. these forces cannot stop the contraction. to about 106 g/cm3 ! The star is a white dwarf .

They are always members of a binary star system. Quasars sit at the center of some galaxies. For poorly understood reasons. One possible result of a Type II supernova is a neutron star. It was visible during the day for 23 days and outshone all other stars in the sky for several weeks. Even degenerate neutron pressure cannot stop the contraction if the star is too massive. If the Earth should happen to be in the direction of the cone periodically as the star rotates. A typical neutron star has a radius of 10 km and a density of 1014 g/cm3 ! Neutron stars have manifested themselves in two ways. The neutron star is a pulsar . Eventually a supernova of another kind. We close our catalog of remarkable astronomical objects with quasars. then the star will appear to pulse at the frequency of rotation of the star.6.4 solar masses. there can be a small region of the star that emits radio and optical frequency radiation in a rather narrow cone. mostly distant. A typical quasar emits 100 times the energy of our entire Milky Way galaxy from a region 1017 times smaller! 51 . There are also neutron stars which pulse strongly in X-rays. degenerate electron pressure cannot stop the contraction. fantastic pressures force most of the electrons to combine with protons to form neutrons. In a neutron star. Today we see the material blown from the star as the Crab Nebula.1 Stellar Evolution For a star over 1. The outer portions of the star blow into interstellar space. We discuss black holes in Sec. discovered in 1963. There is a pulsar at the center of the crab nebula. We discuss these systems in Sec. The ﬁrst pulsar was discovered in 1968. A supernova was recorded in China in 1054. A massive star can collapse to a black hole. 3. called Type II.3.up to nearly 1000 times a second. 3.6. Degenerate neutron pressure prevents the star from collapsing further. They often spin rapidly . many are now known. occurs.

This will enable us in the next section to compare the predictions of general reltivity with observations of the motion of planets and light in our solar system. (2.3. Thus ∂ν/∂t = 0. We ﬁrst set Rtr = −2 (∂ν/∂t)/r = 0 .1. dtdθ. dθdφ.) • dθ and −dθ produce the same ds.3) This is as far as we can go with symmetry and coordinate changes. Thus none of the terms drdθ. Hint: Think geometrically. Since we are interested only in the spacetime outside the central object. (3.11). (Then rename the ¯ radial coordinate r. These coordinate changes put the metric in a simpler form: ds2 = e2µ dt2 − e2ν dr2 − r2 d Ω 2 . Now Rtt = µ − µ ν + µ Rrr = µ − µ ν + µ Rφφ = −rµ + rν + e 2 2 2ν + 2 µ r−1 − 2 ν r−1 e−4ν −1 e −2ν (3.. u. we use the vacuum ﬁeld equation Eq. Exercise 3. (3. Show that in spherical coordinates the ﬂat spacetime metric Eq. r = ro has the metric of a sphere.1) where from Eq. The coordinate change r = reλ eliminates the e2λ factor. of θ or φ. (3. ν. (3. How can Eq. as do dφ and −dφ. i. We must use the ﬁeld equation to determine µ and ν. (2.14) becomes ds2 = dt2 − dr2 − r2 dΩ 2 . See Appendix 12. They are best calculated with a computer. (1. dΩ 2 = dφ2 + sin2 φ dθ2 is the metric of the unit sphere.2) where µ. and λ are unknown functions of t and r.2 The Schwartzschild Metric 3.) A change in the t coordinate eliminates the dtdr term. but not.1) change in a curved spherically symmetric spacetime? In such a spacetime: • The θ and φ coordinates can have their usual meaning. by symmetry. (As in a curved surface. or dtdφ can appear in the metric. although not necessarily of radius ro . angles can be measured in the usual way in a curved spacetime. • The surface t = to . 52 . We write some of the coeﬃcients as exponentials for convenience.25). (2. The components of R are given by Eq. drdφ.e.28): R = 0.2 The Schwartzschild Metric In this section we obtain the metric for the spacetime around a spherically symmetric object such as the Sun. ν depends only on r.4) . Thus the metric is of the form ds2 = e2µ dt2 − 2udtdr − e2ν dr2 − r2 e2λ d Ω 2 .

Exercise 3. Thus the Schwartzschild metric reduces to the ﬂat spacetime metric. the matter outside the sphere exerts no gravitational inﬂuence inside the sphere. (3.47 km. For the Sun. m = 4. Thus if r 2m.2.5) We shall see later that m = κM. Set Rtt = 0 and Rrr = 0. otherwise there would be a singularity in the metric at r = 0.92 × 10−6 sec = 1.1) are nearly identical. where 2 m is a constant of integration. (3.5) and the ﬂat spacetime metric Eq. Thus if the central object is expanding or 53 . Show that the Schwartzschild metric predicts the sum of the results of Exercises 1. Now Rφφ = 0 implies 2 rν + e2ν − 1 = 0. and subtract: µ + ν = 0.9 and 2.1) and M is the mass of the central object. Thus general 2 relativity gives both the time dilation and gravitational redshift diﬀerences. and nothing more. (3.6) where κ is the Newtonian gravitational constant of Eq.1 for the diﬀerence between the clocks in the 2 2 Hafele-Keating experiment: ∆sa − ∆sg = 1 (va − vg ) + gh ∆t . It is remarkable that the Schwartzschild metric is independent of t even though we did not assume this. (2. or (r e−2ν ) = 1. then the Schwartzschild metric Eq. In this case m = 0. Thus the substitution ¯ dt = eµ+ν dt eliminates the e2(µ+ν) factor. cancel the exponential factor. It is interesting to note that our derivation of the Schwartzschild metric applies to the interior of a hollow sphere surrounded by a spherically symmetric distribution of matter.7) Exercise 3. Show that the time ds measured by a clock at rest at r is related to the coordinate time dt by ds = 2m 1− r 1 2 dt. Since µ + ν = 0. Since ν does not depend on t. (3. (3. then for most purposes r may be considered radial distance and t time measured by slowly moving clocks.3) is now of the form ds2 = 1− 2m r e2(µ+ν) dt2 − 1 − 2m r −1 dr2 − r2 d Ω 2 .3. obtained by Karl Schwartzschild in 1916: Schwartzschild Metric ds2 = 1− 2m r dt2 − 1 − 2m r −1 dr2 − r2 d Ω 2 . (3. Integrate: e−2ν = 1 − 2 m/r. µ + ν is a function only of t.3. If r 2m. The metric Eq. the Schwartzschild metric.2 The Schwartzschild Metric where primes indicate diﬀerentiation with respect to r. We arrive at our solution. neither does m.

(3. Observe them at (to .8) Exercise 3. After taking into account the redshift due to the motion of the rocket. (1. (3. ro ) and (to + ∆to . For light emitted at the surface of the Sun and received at Earth.6) and Eq. By Eqs. We now investigate the gravitational redshift using the Schwartzschild metric. such as supernovae (see Sec. have a larger redshift.8).8) reduces to Eq. then Eq. Such an object cannot emit gravitational radiation – propagated disturbances in the metric caused by matter in motion. Of course the gravitational redshift changed with the changing height of the rocket. the gravitational redshift formula derived in connection with the terrestrial redshift experiment. 1 − 2m/re (3. via radio.35 has been measured in X-rays emitted from the surface of a neutron star. It can carry energy away from its source. Thus ∆to = ∆te . (3. with ∆so deﬁned similarly. Eq. This is diﬃcult to measure but it has been veriﬁed within 7%. 4. (3. 3.2).3. Show that if re 2m and ro − re is small. re ) and (te +∆te .) An accurate measurement of the gravitational redshift was made in 1976 by an atomic clock in a rocket.8). Show that a clock on the Sun will lose 63 sec/year compared to a clock far from the Sun. with one on the ground during the two hour ﬂight of the rocket. and the big bang (see Sec. 3.4. The mathematical description of gravitational radiation and the physical interpretation of the mathematical description are complicated matters beyond the scope of this book. collisions of stars (see Sec. re ). respectively. the worldline of the second pulse is simply a tranlation of the ﬁrst by ∆te . For example. (But see Sec. Note that c = 1 in Eq. 54 . So far none has been detected. Violent astronomical phenomena.5. by Eq.) The diﬃculty in detecting this radiation is its extreme weakness.1). light from the white dwarf companion to Sirius has a gravitational redshift z = 3 × 10−4 . Several detectors have been built and are being built to try to detect gravitational radiation. (2. And a gravitational redshift z = . Since the Schwartzschild metric is time independent. Exercise 3.2 The Schwartzschild Metric contracting in a spherically symmetric way. ro ).5. then the metric cannot change. 3. Let ∆se be the time between the emission events as measured by a clock at re .5). The reading of the clock was compared.7) an observer at ro ﬁnds a redshift z= ∆so −1= ∆se 1 − 2m/ro − 1. Emit pulses of light radially outward from (te . (The corresponding losses for the white dwarf and neutron star just mentioned are 2. Light from a star with the mass of the Sun but a smaller radius will.6 hours and 95 days. (3. the gravitational redshift predicted by general relativity was conﬁrmed within 7 parts in 105 .1) emit gravitational radiation. The radiation travels at the speed of light.8) gives z = 2 × 10−6 .

11): ˙ t (1 − 2m/r) = B ˙ θ r2 = A. They are: 2m/r2 ¨ t+ rt = 0 ˙˙ (3.11) r ¨ 2 ˙˙ ˙ φ + rφ − sin φ cos φ θ2 = 0.11). (3. This is a solution to Eq.3. (3. we must solve the geodesic equations Eq. giving (ds/dp)2 = E.14) and (3.9) and (3.12).16) where E is a constant of integration. (3. (3. Exercise 3. (3. 3.14) (3.) Substitute Eqs.5).2 The Schwartzschild Metric To determine the motion of inertial particles and light in a Schwartzschild spacetime. (3.16) into the Schwartzschild metric Eq. (2. Use Eqs.12) r From spherical symmetry a geodesic lies in a plane.13)-(3. A and B are conserved quantities: A is the angular momentum per unit mass along the geodesic and B is the energy per unit mass. Exercise 3.2. (Cf.5) to show that ds = (1 − 3m/r) 2 dt for a circular orbit at φ = π/2.7. (3. It also shows that light (with ds = 0) can orbit at r = 3m.21). For an inertial particle we may take p = s (see Eq. This is not so for ordinary objects.9) and (3. The values given for E are veriﬁed by substituting Eqs. (3.10) and (3.6.9) 1 − 2m/r m(1 − 2m/r) ˙2 m/r2 ˙ ˙ r+ ¨ t − r2 − r(1 − 2m/r) φ2 + sin2 φ θ2 = 0 (3.) This shows that a clock (with ds > 0) can have a circular orbit only for r > 3m. (3. (2. The circular orbits of the exercise are of course possible only if the central object is inside the sphere r = 3m. Diﬀerentiate Eqs. See Sec. (2.21)) and so E = 1.15) 1 (3. (3. (3. Let it be the plane φ = π/2 .15) to obtain Eqs. Physically.13) where A and B are constants of integration.6. Integrate Eqs.16)) and so E = 0. but it is for a black hole. For light ds = 0 (see Eq.10) ˙ r2 1 − 2m/r ¨ 2 ˙ ˙ ˙ ˙ θ + r θ + 2 cotφ φ θ = 0 (3.15) into Eq.13)-(3.10) and integrate: r2 ˙ A2 B2 + 2 − = −E = 1 − 2m/r r 1 − 2m/r 0 for light −1 for inertial particles. (3. 55 . (Remember that φ = π/2 . Exercise 3.

13)-(3. Since the Newtonian theory applies to this situation. Eqs. ds ≈ dt. in agreement with Eq.18) must coincide. Thus Eq. In Eqs.2 The Schwartzschild Metric The partially integrated geodesic equations Eqs. A = 0.13)-(3.16) becomes 2m dr = ± B2 − 1 + ds r 1 2 .16). • The Ricci tensor involves second derivatives of a metric. (3. We are now able to evaluate the constant m in the Schwartzschild metric Eq.15). (2. (3.3. Diﬀerentiate Eq. (3. They are of second order.5).17) and substitute Eq.17) into the result: m d2 r = − 2. (3. so it might be useful to outline the steps we used to obtain them: • We obtained a general form for the metric of a spherically symmetric spacetime.12) we wrote the geodesic equations for the Schwartzschild metric. Thus m = κM .6). In Eqs. We took several side journeys enroute to the equations. • The geodesic equations involve ﬁrst derivatives of a metric. Eq. (3. • We integrated the second order geodesic equations to obtain ﬁrst order Eqs. (3. • We wrote the vacuum ﬁeld equations for the metric by setting the components of the Ricci tensor to zero.17) the sign chosen according as the motion is outward or inward.5).9)(3. • We solved the vacuum ﬁeld equations to obtain the Schwartzschild metric. By Eq. 2 ds r (3. (3. (3. Consider radial motion of an inertial particle.18) For a distant slowly moving particle.4) we wrote some of the components of this tensor for the spherically symmetric metric. 56 . (3. (3. Eq. (3.2).1) and (3. (3. (3.16) will be the starting point for the study of the motion of planets and light in the solar system in the next section.

20) and (3. although Einstein’s theory is conceptually entirely diﬀerent from Newton’s.22): −1 u = rp (1 + e) −1 1 + e cos θ − θp + 3mθ rp (1 + e) . cos β + α sin β ≈ cos β cos α + sin β sin α = cos(β − α).3. and divide by 2A2 du/dθ: d2 u + u = mA−2 E + 3mu2 . Mercury has the largest ratio of the terms on the right side of Eq. Use this approximation in Eq. it gives nearly the same predictions for the planets. (3. We ﬁrst solve the geodesic equations for the planets. Set u = 1/r.3 The Solar System Tests In this section we compare the predictions of general relativity with observations of the motion of planets and light in our solar system. substitute Eqs.19).5) derived from the Newtonian theory except for the term 3mu2 .19).21) into the right side of Eq. (3. multiply by 1 − 2mu. (3. (3. which stays small because of the factor mrp . Thus. As a ﬁrst approximation.3 The Solar System Tests 3. Perihelion Advance. Show that it is less than 10−7 . (3. and light retardation. r= ˙ dr du ˙ du du θ = −u−2 A r−2 = −A .19) and solve: −1 u = rp (1 + e) −1 −1 1 + e cos(θ − θp ) + 3mrp (1 + e) −2 −1 θ sin(θ − θp ) (3.8. (3. dθ2 (3. We discuss three general relativistic eﬀects: perihelion advance.21) To obtain a more accurate solution of Eq.16). Exercise 3. (A. (3. (3.19) For an inertial particle E = 1.23) 57 .22) −2 + mrp (1 + e) 1 3 + 2 e2 [ 3 − cos(2(θ − θp ))] . solve Eq.19) without the small 3mu2 term: u = mA−2 [1 + e cos(θ − θp )] . light deﬂection.20): −1 mA−2 = rp (1 + e)−1 (3. For the planets this term is very small.15).20) This is the equation of an ellipse with eccentricity e and perihelion (point of closest approach to the Sun) at θ = θp . du dθ dθ dθ Substitute this into Eq. This is necessary for any theory of gravity. (3.19) is identical to the equation of motion Eq. For small α. and Eq. Among the planets. (3. (3. Set θ = θp in Eq. By Eq. diﬀerentiate with respect to θ. as Newton’s theory is very accurate for the planets. −2 Drop the last term. (3.

Set E = 0 in Eq. (3. The explanation of the discrepancy was the ﬁrst observational veriﬁcation of general relativity.9.3. (2. (The Newtonian equation Eq. He conﬁrmed the prediction of general relativity within about 20%. Today this prediction of general relativity is veriﬁed within . dθ2 (3. but not by much.19) for light to give the geodesic equation Fig.0 arcsecond/century. The 43 arcsec was the only known discrepancy between Newton’s theory and observation when Einstein published the general theory of relativity. this may be considered the equation of an ellipse with perihelion advanced by 6mπ/[ rp (1 + e) ]/revolution.5%.1. Substitute this into the right side of Eq. cos(π + δ) = − cos δ ≈ −1 to obtain the observed deﬂection angle δ = 4m/rp . Stars can be seen near the Sun only during a solar eclipse. p 2 Set u = 0 (r = ∞. Later eclipse observations have improved the accuracy. Show that for Mercury the predicted perihelion advance is 43. Show that for the Sun. δ = 1. Consider light which grazes the Sun at θ = 0 when u = up = 1/rp .24) and solve to obtain a better approximation: u = up cos θ + 1 mu2 (3 − cos 2θ). 3. light passing near the Sun will be deﬂected.01% using radio waves emitted by a quasar. Since 1859 astronomers explained all but about 43 arcsec as due to slight gravitational eﬀects of other planets.) Exercise 3.1) predicts half of this deﬂection. Sir Arthur Stanley Eddington organized an expedition to try to detect the deﬂection during the eclipse of 1919. 3.10. Light Deﬂection.24) Ignore the small 3mu2 term to obtain a straight line as an approximate solution: u = up cos θ. d2 u + u = 3mu2 . The perihelion of Mercury is observed to advance ∼ 500 arcsec/century. In 1995 the predicted deﬂection was veriﬁed within . several radio telescopes working together can measure angles more accurately than optical 58 . Radio waves have two advantages.75 arcsec. According to general relativity. First. Exercise 3. (3. See Fig.1: Deﬂection of light by the sun. θ = ±(π/2 + δ/2) and approximate cos(π/2 + δ/2) = − sin δ/2 ≈ −δ/2.3 The Solar System Tests Since 3m/[ rp (1 + e) ] is small. Less than an arcminute per century! This is the remarkable accuracy of Newton’s theory.

substitute the above two equations into the result. 1 (3. Second. then the quasar appears as one or several arcs around the galaxy.2 shows the cause of the double image of the quasar. an eclipse is not necessary. reﬂected oﬀ a planet. Light Retardation. 3. But in many cases this is nearly so.16): ˙ 2 A2 B −2 = rp 1 − −1 . use Eq. Radar can be sent from Earth.3 The Solar System Tests telescopes. are in fact the same quasar! Fig. The galaxy is a gravitational lens.25) In Appendix 13 we approximate the integral: 2 2 t = (re − rp ) 2 + 2m ln(2re /rp ) + m.3: Radar echo delay. and integrate: re 1− t= rp 2 2m −2 r rp 2 1−2m/r 1−2m/rp r 1− 1 dr. Divide Eq. To calculate the time t for light to go Fig. 6 arcsec apart. (3. and detected upon its return to Earth. and so the deﬂection can be measured. If this is done as the planet is about to pass behind the Sun. 2m rp −1 . (3. And in at least one case the quasar appears as a ring around the galaxy! A gravitational lens can also brighten a distant quasar by up to 100 times. from Earth to Sun. (3. 3. Many multiple (up to 7) image quasars are now known. (3. See Fig. Light arriving from a quasar at an angle of 90◦ from the sun is deﬂected by 4 × 10−3 arcsec by the sun. Two quasars. 3.3. as radio sources can be detected during the day. this to occur. the radar is slowed by the Sun’s gravity. separate the variables. 3.2: Gravitational lens.16) by B 2 .3.14): dr dt dr 2m dr = = B 1− dp dt dp dt r Set r = rp and r|r=rp = 0 in Eq. A spectacular example of the gravitational deﬂection of light was discovered in 1979. The accuracy of such measurements is now better than 10−4 arcsec. enabling astronomers to study them better.26) 59 . It is not necessary that the quasar be exactly centered behind the galaxy for Fig. then according to general relativity.

The lunar laser experiment has conﬁrmed this to 1%. This is the geodetic eﬀect. In Fig. 3. Fig. that of the northern is 5 arcsec. But this is more than compensated for by the gravitational delay of the light of the southern image passing closer to the lensing galaxy. by the Pythagorean theorem. The other terms represent a lengthening of this ﬂat spacetime t. One diﬃculty in performing the experiment lies in determining the positions of the planets in terms of the r coordinate of the Schwartzschild metric accurately enough to calculate the ﬂat spacetime time. 3.02 arcsec/yr. The geodetic eﬀect predicts a change of the axis of rotation of ∼ .5 a vector on a sphere is moved parallel to itself (i. 3. The vector returns to A with its original direction. The Earth-Moon system is a “gyroscope” orbiting the Sun. The same is true in local inertial frames in a curved spacetime. But over a worldline the orientation of the axis with respect to the distant stars can change. This is a manifestation of curvature. 3. is motivated in Figs.3. The southern image of the quasar is 1 arcsec from the lensing galaxy.4 × 10−4 sec. We see variation in the southern image of the original double quasar about 1 1 years after the same 2 variation of the northern image. In Fig. a vector in a plane is moved parallel to itself from A around a closed curve made up of geodesics. double to include the return trip. (re − rp ) 2 . returns to A rotated through angle α.5: A parallel transported vector returns to A with its original direction.. For example. a gyroscope in a circular orbit at coordinate r in a Schwartzschild spacetime will precess about an axis 1 normal to the plane of the orbit by an angle 2π[1 − (1 − 3m/r) 2 ]/revolution in the direction of the orbit.5.4. parallel to itself in local planar frames). 60 .4 and 3.e. 3.3 The Solar System Tests 2 2 The ﬁrst term. from A around a closed curve made of geodesics. Add analogous delay terms for the path from Sun to Planet. and ﬁnd for Mercury a delay of 2. The vector returns to A rotated through the angle α.4: A parallel transported vector Fig. This diﬃculty and others have been overcome and the prediction veriﬁed within 5%.1% Quasars can vary in brightness on a timescale of months. Signals from the Viking spacecraft on Mars have conﬁrmed this eﬀect within . causing a delay in the northern image. Thus the light of the southern image travels less distance than that of the southern. 1 Another eﬀect of general relativity. Willem deSitter’s geodetic eﬀect. is. the time required for light to travel in a straight line in a ﬂat spacetime from Earth to Sun. The axis of rotation of a gyroscope moves parallel to itself in inertial frames in a ﬂat spacetime .

r = const in a Kerr spacetime do not have the metric of a sphere. All of the gravitational eﬀects that we have discussed until now are eﬀects of the mass of an object.28) r3 m 1 2 + a. Show that if the central object is not spinning. Gravitomagnetism is a gravitational eﬀect of the motion of an object. as in the Schwartzschild metric. This is a shift of ∼ 1 m/yr in the intersection of the orbital and equatorial planes.27) − Σ dφ2 − r2 + a2 + where m = κM . describes the spacetime outside a rotating object which is symmetric around its axis of rotation and unchanging in time: Kerr Metric ds2 = 1− 2mr Σ dt2 + 4mar sin2 φ Σ dt dθ − dr2 Σ ∆ 2ma2 r sin2 φ Σ sin2 φ dθ2 . a spinning object.29) 61 . the object cannot be rotating. (3.4 Kerr Spacetimes The Schwartzschild metric describes the spacetime around a spherically symmetric object. by virtue of its orbital motion. The Kerr metric predicts that the axis of a satellite in polar orbit at two Earth radii will change by ∼ . (3. The Lense-Thirring or frame dragging gravitomagnetic eﬀect causes a change in the orientation of the axis of rotation of a small spinning object in the spacetime around a large spinning object.4 Kerr Spacetimes 3. This is the ﬁrst direct measurement of gravitomagnetism.11.03 arcsec/yr. An Earth satellite is. We derive this gravitomagnetic clock eﬀect from the r-geodesic equation for circular equatorial orbits in the Kerr metric: r3 ˙ ˙˙ t 2 − 2atθ + a2 − m Solve for dt/dθ: dt dθ =± ± ˙ θ2 = 0. a = J/M . Note the surfaces t = const. ∆ = r2 − 2mr + a2 . Exercise 3. Observations of two LAGEOS satellites in non-polar orbits have probably detected the Lense-Thirring eﬀect within 30% of the prediction of general relativity. then the Kerr metric Eq. the angular momentum per unit mass of the central object. Its name refers to an analogy with electromagnetism: an electric charge at rest creates an electric ﬁeld. Satellites in the same circular equatorial orbit in a Kerr spacetime but with opposite directions take diﬀerent coordinate times ∆t± for one complete orbit θ → θ ± 2π.3. (3. while a moving charge also creates a magnetic ﬁeld. discovered only in 1963 by Roy Kerr. whose axis of rotation is perpendicular to its orbital plane. Σ = r2 + a2 cos2 φ .5). and the axis of rotation is the z-axis. (3. In particular. (3. The Kerr metric.27) reduces to the Schwartzschild metric Eq.

Show that for the Earth.3. an orbit in the sense of a takes longer than an opposite orbit.4 Kerr Spacetimes where ± corresponds to the orbits θ → θ ± 2π. ∆t+ − ∆t− = 2 × 10−7 sec. (3.12. This has not been measured.29) over one orbit by multiplying the right side of the 1 equation by ±2π: ∆t± = 2π(r3 /m) 2 ± 2πa. Exercise 3. Integrate Eq. 62 . Then ∆t+ − ∆t− = 4πa.

This is the best empirical evidence for the existence of gravitational radiation. Both stars are 1.2◦ /year – 35. the stars come closer together. Several changing relativistic eﬀects aﬀecting the arrival times at Earth of the pulsar’s pulses must be taken into account while measuring a change in the period: the gravitational redshift of the rate of clocks on Earth as the Earth’s distance from the Sun changes over the course of the year. rather than the vacuum ﬁeld equation.4 solar mass neutron stars. Imagine! By great good fortune. the gravitational redshift of the time between the pulses as the pulsar moves in its highly elliptical (eccentricity = .5 The Binary Pulsar 3. The orbital period just before the collision will be of the order of 10−3 sec! 63 .000 times that of Mercury. Calculations predict that the orbital period of the system should decrease by 10−4 sec/year due to energy loss from the system due to gravitational radiation. Both are present in a remarkable binary star system discovered in 1974. a “slowing” of the pulsar clock due to its motion. analogous to the delay measured in the radar echo delay experiment.5%. By 1978 the period was known to decrease by about the amount predicted by general relativity. the diﬀerences between Einstein’s and Newton’s theories of gravity are exceedingly small in the solar system. More complicated techniques using the full ﬁeld equation. They will collide in about 300 million years.6) orbit. periastron is the generic term. The periastron advance of the pulsar is 4.) The Schwartzschild metric cannot predict this because neither star is an inertial particle in the gravity of the other. and the two stars are about a solar diameter apart. Large diﬀerences require strong gravity and/or large velocities. The prediction is now conﬁrmed within . The pulsar’s pulses serve as the ticking of a very accurate clock in the system. Their orbital period is eight hours. (Perihelion refers to the Sun. Several relativistic eﬀects have been measured by analyzing the arrival times of the pulses at Earth.5 The Binary Pulsar As we have seen. As the binary system loses energy by gravitational radiation. one of the neutron stars is a pulsar spinning at 17 revolutions/sec. These measurements provide the most probing tests of general relativity to date. earning the system the title “Nature’s gift to relativists”.3. must be used. and a delay in the pulses.

the Schwartzschild radius is 3 km and for the Earth it is . the Schwartzschild metric Eq. Exercise 3.9 cm. Use Eqs. so that r−2m is small. To see a coordinate singularity in a more familiar setting.6 Black Holes The Schwartzschild metric Eq. (3. An observer crossing the horizon would not notice anything special. The object is then a black hole. a true spacetime singularity. The only singularity is at r = 0. For the Sun. (3. a solution of the vacuum ﬁeld equation valid only outside the central body. neither matter nor light can escape a Schwartzschild black hole! The same calculation shows that it takes an inﬁnite coordinate time ∆t for matter or light to enter a black hole.3. Introduce Painlev` coordinates with the coordinate change e ¯ dt = dt + (2mr) 2 dr r − 2m 1 in the Schwartzschild metric (then drop the bar) to obtain the metric ds = dt − 2 2 dr + 2m r 1 2 2 dt − r2 d Ω 2 .6 Black Holes 3.14) and (3.13. but there is no geometric singularity. Thus the total coordinate time ∆t satisﬁes r2 ∆t ≥ 2m 2m 1− r −1 r2 dr = 2m 1+ 2m r − 2m dr = ∞. Suppose a material particle or pulse of light moves outward (not necessarily radially) from r1 = 2m to r2 . The r = 2m surface is called the (event) horizon.14. substitute u = 1/r in the ﬂat surface polar coordinate metric ds2 = dr2 + r2 dθ2 to obtain ds2 = u−4 du2 + u−2 dθ2 . 64 .5) has a singularity at the Schwartzschild radius r = 2m: grr = ∞ there. there is no singularity in the spacetime.17) and approximate to show that dr/dt = (2m − r)/2m.5) shows that dt ≥ (1 − 2m/r)−1 dr. This is well inside these bodies. We may inquire however about the properties of an object which is inside its Schwartzschild radius. Exercise 3. Since ds2 ≥ 0 and dΩ2 ≥ 0 . (3. There a coordinate singularity at the origin. and so the singularity has no eﬀect on the Schwartzschild metric. Suppose an inertial object falling radially toward a black hole is close to the Schwartzschild radius.13 shows that the singularity at the Schwartzschild radius is only a coordinate singularity. Exercise 3. Integrate and show that r − 2m decreases exponentially in t.

The proofs are more diﬃcult than for the Schwartzschild metric. Due to conservation of angular momentum. and motion must be in the direction of the rotation! 65 . and gtt = 0 on the surface Σ = 2mr. Since ds2 ≥ 0 for both particles and light. (3. (3. the redshift of an object approaching a Schwartzschild black hole increases to inﬁnity as the object approaches the Schwartzschild radius. In particular. never to return. |dr/ds| increases as r decreases. As with the Schwartzschild horizon. dr2 .. i. is no singularity in the spacetime on either surface. a(dθ/dt) > 0.8).15). Show that the static limit has equation r = m + (m2 − a2 cos2 φ) 2 .17). a. a rotating star collapsed to a black hole will spin very rapidly. dφ2 . its brightness (rate of emission of light) will dim to zero and it will eﬀectively disappear. Fig.15. 1 b. (2. observe that by the metric postulate Eq. For a Kerr black hole. In the ergosphere. the dt2 . even though it is possible to reach it in ﬁnite proper time. Most stars rotate.6: The horizon and static limit of a Kerr black hole. it is impossible to escape the Kerr horizon.6 Black Holes According to Eq. grr = ∞ and gtt = 0 on the surface r = 2m. The crescent shaped region between the horizon and the static limit is the ergosphere. and dθ2 terms in the Kerr metric Eq. 3. and the outer surface Σ = 2mr (called the static limit). the horizon. In the ergosphere it is impossible to be at rest. A distant observer will see the rate of all physical processes on the object slow to zero as it approaches the Schwartzschild radius. the dtdθ term must be positive.6 shows the axis of rotation of the Kerr black hole. the inner surface ∆ = 0 (called the horizon). in a ﬁnite proper time ∆s measured by its clock. 3. 1 Exercise 3. On the other hand.3. By Eq.e. Thus the object will cross the Schwartzschild radius. We have seen that there is no singularity in the spacetime there. grr = ∞ on the surface ∆ = 0. There Fig. dr/ds is the rate of change of r measured by a clock carried by the object.27) are negative. For a Schwartzschild black hole. (3. We must use the Kerr metric to describe a rotating black hole. an inertial object radially approaching a Schwartzschild black hole will cross the Schwartzschild radius in a ﬁnite time according to a clock it carries! To see this. They are not given here. Show that the horizon is at r = m + (m2 − a2 ) 2 .

6 Black Holes General relativity allows black holes. We saw in Sec. 3. gas pulled from the normal star and heated violently on its way to the compact companion causes the X-rays. No known force can stabilize a supernova remnant larger than ∼ 3 solar masses. closest approach to the source 17 light days. During “normal” periods. Presumably this is because the companions are black holes. the companion’s mass is small enough to be a neutron star. if not most. Several black holes of a few solar masses are known in our galaxy. white dwarf stars. There is a star orbiting the source with eccentricity . They are members of X-ray emitting binary star systems in which a normal star is orbiting with a compact companion.7 million solar masses. and neutron stars are stabilized by diﬀerent forces. A black hole at the center seems to be the only possibility. Systems with companions over three solar masses never exhibit such periods. and supermassive black holes. 66 . the Sun.87.1 that the Earth. the companion is well over the three solar mass limit for neutron stars. and is therefore believed to be a black hole. of a few solar masses. Some of the systems with a neutron star companion exhibit periods of much stronger X-ray emission. Sometimes the cause is a thermonuclear explosion of material accumulated on the surface. but do they exist in Nature? The answer is almost certainly “yes”. of 106 − 1010 solar masses. which do not have a surface. Since black holes emit no radiation. Sometimes the cause is gas suddenly crashing onto the surface of the neutron star. and with a period of 15 years. At least two varieties are known: solar mass black holes. Quasars are apparently powered by a massive black hole at the center of a galaxy. Many. Orbital data show that in some of these systems. galaxies have a supermassive black hole at their center. they must be observed indirectly. X-ray observations of several galactic nuclei indicate that matter orbiting the nuclei at 3-10 Schwartzschild radii at 10% of the speed of light.3. This implies that the source is 3. A strong radio source at the center of our Milky Way galaxy has a radius of about one light hour. the remnant will collapse to a black hole. But in others.

except a few nearby. But they are not at rest. The Milky Way is about 100.1) holds. other galaxies recede. proFig. Its value is 21 (km/sec)/million light years. (4. This is an audacious move.1: Balloon analogy to vides a very instructive analogy. not painted on.1) is Hubble’s law . The Sun is but one star in 200 billion or so bound together gravitationally to form our galaxy. 4. Eq. Galaxies are inertial objects. the study of the universe as a whole. An expanding spherical balloon. It is but one galaxy in tens of billions in the visible universe. H is Hubble’s constant. We see the balloon expanding in an already existing three dimensional space. we use it to construct a model of the spacetime of the entire universe. and Eq. the Milky Way.1. (4. within 5%. From the viewpoint of every galaxy on the balloon. but the model is very successful. The nearest major galaxies are a few million light years away. 4. All galaxies. but surface dwellers are not aware of this third spatial dimension. as we shall see. See Fig. galaxies separate not because they 67 . our universe’s three dimensional space.Chapter 4 Cosmological Spacetimes 4.000 light years across. because they do not expand with the universe. Since general relativity is our best theory of space and time. each representing a galaxy. The balloon’s two dimensional surface is the analog of the universe. The galaxies are glued. recede from us with a velocity v proportional to their distance d from us: v = Hd. (4. on which bits of paper are glued.1) This is the “expansion of the universe”. We begin with a description of the universe as seen from Earth.1 Our Universe I This chapter is devoted to cosmology. For both surface dwellers in their universe and we in ours.

1 Our Universe I are moving apart in a static ﬁxed space. as follows. z = fe /fo − 1 (see Eq. 68 . the universe contracts. and thus at an exceedingly high temperature. exhibits a redshift z > 0. But they can be measured indirectly. By deﬁnition. its big bang. The absolute luminosity of various types of stars and galaxies has been approximately determined. Thus they all arrive here at the same time. to moderate distances we can ignore relativistic eﬀects and determine v using the approximate Doppler formula z = v from Exercise 1. We cannot verify Eq. the result of the expansion of space. it occurred everywhere. we can use z (which is directly and accurately measurable). Due to the ﬁnite speed of light.6b. Light from all galaxies. ıtre There is no “site” of the big bang. 4πd 2 (4. Imagine it expanding from a point. The recession of the galaxies that we see today is the continuation of the explosion. Thus we say. When we look out in space.. The idea of a hot early universe preceding today’s expansion was ﬁrst proposed by Georges Lemaˆ in 1927.g. Reversing time. (1. (4. Since z increases with d. e. Fig. then z is directly and precisely measurable. Substituting z = v into Eq.2) Thus the distance to the object can be determined from a measurement of its luminosity (or the luminosity of a resolved star in it) if its absolute luminosity is known. The distance d to objects at moderate distances can be determined from their luminosity (energy received at Earth/unit time/unit area). the matter in the universe was extremely compressed.1 directly because neither d nor v can be measured directly. called the big bang. Our balloon again provides an analogy. except a few nearby. valid to moderate distances. that a galaxy is “at” z = 2. From Eq. accompanied by exceedingly intense electromagnetic radiation. Continuing into the past. we see a galaxy not as it is today.1) galaxies approach us with a velocity proportional to their distance.7 billion years ago. At a later time. but because space itself is expanding. as a proxy for d (which is not).1) gives z = Hd. However. This approximation to Hubble’s law was established in 1929 by Edwin Hubble.7)). It occurred 13. then its observed luminosity at distance d is its absolute luminosity L (energy emitted/unit time) divided by the area of a sphere of radius d: = L . Matter and space were taking part in an explosion. but as it was when the light we detect from it was emitted. 4. If the object is near enough so that relativistic eﬀects can be ignored. (4. allowing an approximate determination of their distances. The big bang is the origin of our universe. to moderate distances. there is no speciﬁc place on its surface where the big bang occurred. If fe is known (for example if the light is a known spectral line). Eq. carrying the galaxies with it.12) shows that the redshifts are not Doppler redshifts. The velocity v of galaxies to moderate distances can be determined from their redshift. (4. they are ex:pansion redshifts. we look back in time! This allows us to study the universe at earlier times.4.

Explain this as due to the Earth moving toward Leo at 370 km/sec with respect to an isotropic CBR. i. Telescopes today see some galaxies as they were when the universe was less than 1 billion years old. The big bang is part of the most remarkable generalization of science: The universe had an origin and is evolving at many interrelated levels. to be incorporated into new stars. Other elements constitute only a small fraction of the mass of the universe. Observations conﬁrm the prediction. The theory predicts that from a few seconds to a few minutes after the big bang. their planets. biological evolution. planetary evolution. should still be with us today. and cultural evolution. stellar evolution. In this sense we have discovered the absolute motion of the Earth. with traces of some other light elements. allowing nuclei and electrons to condense into neutral atoms. the age of the universe (in 10 toward us ∼ 380. The CBR has a 2. where θ is the angle from the constellation Leo. Further evidence for a big bang comes from the theory of big bang nucleosynthesis.1. After correcting for the dipole anisotropy. 4. They are mostly formed in stars and are strewn into interstellar space by supernovae and by other means. the CBR is isotropic to 1 part in 10. about 3000 K.e. and any life which may arise on the planets. 4. 000 years after the big years) when the light we see from it bang. As a ﬁrst approximation the CBR is isotropic.7 K blackbody spec. quite by accident and unaware of the prediction.4. nuclei formed.4 we shall see that it contains a wealth of cosmological information.Fig. almost all hydrogen (75% by mass) and helium-4 (25%). The radiation started its journey vs. The levels include galactic evolution. This cosmic background radiation (CBR) provides strong evidence that a big bang actually occurred. which made the universe transparent to the radiation.0012 cos θ.. discovered microwave radiation with the proper characteristics. when the universe had cooled to was emitted. In 1965 Arno Penzias and Robert Wilson. 69 .21). This is a spectacular success of the big bang theory. (4. In Sec. in 1977 astronomers measured a small dipole anisotropy: a relative redshift in the CBR of z = −. much diluted and cooled by the expansion of the universe.2: The redshift z of a galaxy 9 trum.000. In 1948 Ralph Alpher and Robert Herman predicted that the intense electromagnetic radiation from the big bang.2. However.1 Our Universe I 4. which is the graph of Eq. The relative redshift is modulated by the orbital motion of the Earth around the Sun. Exercise 4. spherically symmetric about us. illustrates this.

2 Robertson-Walker Spacetimes 4. Exercise 4.10). Choose any galaxy as the spatial origin. (4. φ. The coordinates (r. We ﬁrst set up a coordinate system (t. Redshift surveys of hundreds of thousands of galaxies show that the universe is homogeneous in space. Distances between the glued on galaxies increase. Derive the analog of Eq. (4. where H = S (t)/S(t).1) for the balloon. 70 . r. As the universe expands. but their comoving (φ. If the radius of the balloon at time t is S(t). the density ρ of matter decreases. then according to Eq. like φ and θ. Isotropy ensures that all galaxies at a given r at a given t are at the same distance from the origin. θ) is ds2 = S 2 (t) dr2 + r2 dθ2 (1 + r2 /4) 2 . (2. Deﬁne the r coordinate at t0 to be any quantity that increases with increasing distance from the origin. The t coordinate is the (proper) time measured by these clocks. θ) are called comoving. Galaxies are distributed isotropically. θ) coordinates change. does not change. Substitute sinφ = r/(1 + r2 /4) in Eq. Choose any time t0 . Let d be the distance between two galaxies on the balloon at time t. then the balloon expands. distances between the glued on galaxies increase with S(t) .2 Robertson-Walker Spacetimes The CBR is highly isotropic. but because their t coordinate changes.4. Galaxies in diﬀerent directions but at the same distance are assigned the same r. and let v be the speed with which they are separating. which is what we are interested in.3). θ) coordinates do not change. Show that the metric with respect to the comoving coordinates (r.3) If S(t) increases. which changes the metric. but their comoving (r. (4. Isotropy demands that the φ and θ coordinates of other galaxies do not change in time. (4. θ) for the universe.3. Place a clock at rest in every galaxy and set it to some agreed time when some agreed value of ρ is observed at the galaxy.11) the balloon’s metric at time t is ds2 = S 2 (t) d Ω 2 . θ) coordinates do not change. (4. Exercise 4. Our general relativistic model ignores the clumping of matter into galaxies by assuming an exactly isotropic and homogeneous continuous distribution of inertial matter in the universe. φ. φ. Distances between galaxies change in time not because their (r. Now deﬁne the r coordinate of galaxies at times other than t0 by requiring that r. This should not aﬀect the large scale structure of the universe.4) Again. Show that v = Hd.2. Hint: See the derivation of Eq. Our balloon provides an analogy.

the universe has ﬁnite volume! The surface of our balloon provides an analogy.6) this metric measures physical distances. i.5) gives dt = S(t) dσ.e. By deﬁnition. (4. The substitution ρ = 2 tanh ξ shows that V (2) = ∞.5) ≡ dt2 − S 2 (t) dσ 2 . The metric of points at coordinate radius r in the Robertson-Walker metric Eq. then r is restricted to 0 ≤ r < 2. The substitution ρ = 2 tan ξ shows that V (∞) = 2π 2 S 3 (t).4. Walker in the mid 1930’s. 0. If k = 0 (zero spatial curvature).. and positively curved respectively. If k = 1 (positive spatial curvature). 1 for negatively curved. but then decreases to zero as r → ∞. (4. ±1) (4. We will use it in Sec. P.4). 0 ≤ r < ∞ and the universe has inﬁnite volume. the distance is proportional to S(t). which is thus a scale factor for the universe.6) This is the distance that would be measured by a rigid rod. 4. The ﬁeld equation is not used. Eq. The metric was discovered independently by H. the physical distance between the galaxies is S(t) dσ. Circumferences of circles of increasing radius centered at the North pole increase until the equator is reached and then decrease to zero. It is ﬁnite and has positive curvature.7) V (r) = S(t) 0 A(ρ) dρ . 1 + kρ2 /4 (4. (4. the universe has inﬁnite volume. Robertson and E. The elementary but somewhat involved derivation is given in Appendix 14. Emit a pulse of light from a galaxy to a neighboring galaxy. W. t is measured by clocks at rest in galaxies. The constant k is the curvature of the spatial metric dσ: k = −1. (4.5).8) If k = −1 (negative spatial curvature). Since c = 1 in local inertial frames. or by the Robertson-Walker metric Eq. Thus the sphere has surface area A(r) = and volume r 4πr2 S 2 (t) (1 + k r2 /4)2 (4. Since dσ does not change in time. (4.5) is that of a sphere of radius S(t) r/(1 + k r2 /4). then A(r) increases as r increases from 0 to 2. ﬂat. It should be compared with the metric in Eq. Since ds = 0 for light. (4.2 Robertson-Walker Spacetimes The metric of an isotropic universe. is: Robertson-Walker Metric ds2 = dt2 − S 2 (t) dr2 + r2 d Ω 2 (1 + k r2 /4) 2 (k = 0. From Eq. Thus dt is the elapsed time in a local inertial frame in which the emitting galaxy is at rest. after suitably choosing the comoving r coordinate. 71 .4 to determine S(t).

3 The Expansion Redshift 4. S(t) 1 + k r2 /4 Integrate this over both light worldlines to give to re to +∆to (4. S(t) Subtract to te +∆te from both ends to give te +∆ te to +∆to dt = S(t) te to dt . Since ds = 0 for light. S(t) (4. then v = 1 . 0). re ) and received by us at events (to . where H= S (t) . Galaxies beyond this distance. by Eq.5) gives dt dr = . What about the rule “nothing can move faster than light”? In special relativity the rule applies to objects moving in an inertial frame. (4.10) From v = Hd we see that if d = 1/H ≈ 13. S(t) 72 . Diﬀerentiate to give the velocity of the galaxy with respect to the Earth: re v = S (t) 0 dσ. Eq. d = S(t) 0 dσ. (4. and in general relativity to objects moving in a local inertial frame. re ) and (te + ∆te .9) This is the distance that would be obtained by adding the lengths of small rigid rods laid end to end between Earth and the galaxy at time t. Eq.1). at z ≈ 1. Divide to give Hubble’s law v = Hd. (4. Suppose that light signals are emitted toward us at events (te .8 × 109 light years.6). 0) and (to + ∆to .11) dt = S(t) te 0 dr = 1 + k r2 /4 te +∆te dt .3 The Expansion Redshift re At time t the distance from Earth at r = 0 to a galaxy at r = re is. the speed of light.4. are receding from us faster than the speed of light. (4.5 . There is no local inertial frame containing both the Earth and the galaxy. We now obtain a relationship between the expansion redshift of light and the size of the universe when the light was emitted.

12). fe = (z + 1)fo . According to Eq. (4. According to Planck’s law.6) it would appear to proceed at 1/11 of the rate of the same process on Earth. T (0) (4. Since S = 0 at the big bang.25. For example. From Eq.12) the galaxy is 11 times farther away from us today than when it emitted the light we detect from it! If we could observe a physical process on the galaxy from Earth.5 appeared to last about 1 + . The faint galaxy is magniﬁed 25-100 times by a gravitational lensing galaxy at z = . this becomes ∆to ∆te = .6) to obtain the desired relation: z+1= S(to ) . T (z) . (1.13) The observed spectrum is also blackbody.7 K. then by Eq. at temperature T (0). A galaxy at redshift z ≈ 10 is the most distant known. a supernova at z = . As stated in Sec. (4. Several distant supernovae have exhibited this eﬀect. S(te ) S(to ) Use Eq.13) predicts this: T (3) = (3 + 1)T (0) ≈ 11K. Analysis of spectral lines in a quasar at z = 3 shows that they were excited by an ∼ 11 K CBR. its redshift z ≈ 3000K/2.4. where from Eq. Thus from Eq. The light we see from the galaxy was emitted ∼ 470 million years after the big bang.12) This wonderfully simple formula directly relates the expansion redshift to the universe’s expansion. (4. Radiation emitted at fe is received by us redshifted to fo . Suppose that an object at distance z emits blackbody radiation at temperature T (z). the object radiates at frequency fe with intensity proportional to 3 fe .13). (1.7K ≈ 1100. hfe /kT (z) − 1 e where h is Planck’s constant and k is Boltzmann’s constant. S(te ) (4. 4. 73 .5 times longer than similar supernovae nearby. Thus the intensity of the received radiation is proportional to 3 fo e hfo /kT (0) − 1 where T (0) is deﬁned by z+1= . the CBR was emitted at ≈ 3000 K and is observed today at 2.7). the big bang is at z = ∞.1. (1. Eq. the universe was then about 1/(1 + 3) = 25% of its present size.3 The Expansion Redshift If ∆te and ∆to are small.5 = 1. (4.

the ﬁeld equation was not used in its derivation. the average temperature diﬀerence as a function of angular separation peaks at . Despite the strong evidence for a ﬂat universe. This spatial ﬂatness. The angle is diﬀerent for a curved space.14) This is the most natural change possible to the ﬁeld equation. we have seen the term Λg before. together with observational data. There are two unknowns in the metric: the curvature k and the expansion factor S(t). (4.4 Our Universe II 4.3 and suppose that light travels along the great circles of the sphere. Imagine surface dwellers whose universe is the surface of the sphere in Fig. the density of matter was not quite uniform.4 Our Universe II The Robertson-Walker metric Eq. (2. in the very early universe.8 arcdegree. and expanded the universe by a factor of 1022 ! As fantastic as inﬂation seems. We eliminated it 74 . consider the everyday experience that the angular size of an object decreases with distance. Regions of higher density led to the galaxy clusters we wee today. There is more direct evidence for a ﬂat universe. a few parts in 105 ) in the CBR. but at a slower rate than on a ﬂat surface. Inﬂation predicts a ﬂat universe. By 1990. In 1981 Alan Guth proposed that there was a period of exponential expansion. The simplest solution to this problem is to add a term Λg to the ﬁeld equation: G + Λg = −8πκT. (4. in the ﬁeld equation Eq.4. To understand how curvature can aﬀect angular size.8 arcdegree. providing compelling evidence for a ﬂat universe. the angular size α of an object of length D decreases with distance. we shall see that an application of the ﬁeld equation Eq. (2. lasted perhaps 10−35 sec. size begins to increase. 4.9 these circles are geodesics.27) to the Robertson-Walker metric Eq. On a pseudosphere the angular size decreases at a faster rate than on a ﬂat surface. (4. For an observer at the north pole. In fact. Calculations show that in a ﬂat universe. The object has the same angular size at the two positions in the ﬁgure. In this section we ﬁrst review the compelling evidence that in our universe k = 0. most cosmologists believed that inﬂation occurred. The rate of decrease is diﬀerent for diﬀerently curved spaces. Since then more evidence for inﬂation has accumulated. We then solve the modiﬁed ﬁeld equation to determine S(t). When the CBR was emitted. its angular on a sphere. it explained several puzzling cosmological observations and steadily gained support over the years. 4.3: Angular size And when the object reaches the equator. According to Exercise 2. Observations in 2000 found the peak at . will force us to modify the ﬁeld equation.34). The inﬂation started perhaps 10−34 sec after the big bang.5) is a consequence of isotropy alone. called inﬂation. Fig. leading to tiny temperature anisotropies (∼ 60 mK.5) with k = 0 gives results in conﬂict with observation. Those regions were hotter.

dr = dθ = dφ = 0.16) + 2 SS − ΛS 2 = 0 . . To apply the new ﬁeld equation. We assume that the matter is dust. This conﬁrmed two earlier determinations. (This is certainly reasonable today and certainly unreasonable in the very early universe. The clustering of galaxies is sensitive to the value of some of these parameters. One used spectra of distant galactic nuclei to “directly” measure the density of known forms of matter in the early universe. Some of this matter is in ordinary known forms (neutrons. Today as much as 2/3 of this matter is in vast superhot intergalactic gas clouds.5). The WMAP results determined the value of several cosmological parameters reported in this section. (4. This conﬁrmed earlier estimates based on gravitational lens statistics. 4. electrons. (4. (2. In Sec. Anisotropies in the CBR of a few parts in 105 were measured. θθ components) From Eqs. neutrinos. then its eﬀects in the solar system are negligible. If it is small enough. determination of the parameters.14) for the Robertson-Walker metric Eq. (4. .) Thus T is of the form Eq.15) (4.5) with k = 0 and the T just obtained reduces to two ordinary diﬀerential equations: −3 S 2 S S 2 + Λ = −8πκρ . dt/ds = 1.17) 3 where ρc ≡ 3H 2 /8πκ is called the critical density. and from Eq. but it can dramatically aﬀect cosmological models. The constant Λ is called the cosmological constant. For the comoving matter.3ρc .2 we distributed the matter in the universe into a uniform distribution of density ρ. independent. photons.4. (tt component) (4.10). protons. . 75 . It is gratifying to have excellent agreement between the CBR and deuterium measurements.000 galaxies by the Sloan Digital Sky Survey (SDSS) were also announced in 2003. (4. (rr. 8πκ (4. ). The other inferred the density from the abundance of deuterium created during big bang nucleosynthesis. Thus T tt = ρ is the only nonzero T jk .26). Thus the term represents a gravitational eﬀect of empty space.05ρc . but most is not. as the deuterium was created a few minutes after the big bang the CBR was emitted hundreds of thousands of years later.15) and (4.4 Our Universe II there by assuming that spacetime is ﬂat in the absence of matter. Today ρc ≈ 10−29 g/cm – the equivalent of a few hydrogen atoms per cubic meter. The SDSS results give a second. The ﬁeld equation Eq. Precise measurements of the CBR by the Wilkinson Microwave Anisotropy Probe (WMAP) were announced in 2003. which agree with those of WMAP. we need the energy-momentum tensor T. The measurements show that the known forms of matter constitute . φφ. ρ+ Λ = ρc . The measurements show that today the density of matter is ρ = . Measurements of the positions and redshifts of 200. The deuterium abundance was measured in distant pristine gas clouds backlit by even more distant quasars.

dark matter behaves just as ordinary matter. is constant in time (and space).4: S (in units of S(t0 )) as a ter dominates and at later times dark energy function of t (in 109 years).4. This “it is what is left over in Eq.25ρc . (4. At early times mat. It is therefore called dark matter . For example.18): 3SS 2 − ΛS 3 = 8πκρ0 S 3 (t0 ) . The quantity Λ/8πκ is the density of another mysterious component of our universe.15).15) by −S 3 and use Eq. Density ﬂuctuations of ordinary matter were smoothed out by interactions with electromagnetic radiation.17). Dark energy is intrinsic to space. Gravitationally. (4. 4.17)” argument for its existence is perhaps unsatisfying.18) According to Eq. A quick calculation using Eq.18).4.20) 76 . (4. (4. (4.4 Our Universe II The measurements show that unknown forms of matter constitute .15) and (4. √ (4. From Eq. Show that Λ ≈ 10−35 /sec2 . Thus from Eq. We now solve the ﬁeld equations (4. it literally “comes with the territory”. The separation of the quasar images is so large that dark matter in the cluster must be responsible for the lensing.16) for S. (4. We shall see soon that there is observational evidence for dark energy. This is what we would expect.3ρc = . Galaxies arose from small random density ﬂuctuations in the very early very hot universe.19) Integrate this to solve the ﬁeld equations: 8πκρ0 S (t) = sinh2 Λ 3 3Λ t S 3 (t0 ). the density of dark energy. ρ is inversely proportional to the cube of the scale factor S.16) shows that SS 2 − ΛS 3 /3 = 0. In contrast. Dark matter explains the growth of structure in our universe. Thus the ratio of the dark energy density to the matter density increases with time. and so density ﬂuctuations within it could grow. stars orbiting in the outer part of a galaxy move too rapidly for the density of ordinary matter to hold them in orbit. 2 (4. The existence and approximate abundance of dark matter was ﬁrst inferred from its gravitational eﬀects. Now multiply Eq. (4. Λ/8πκ = ρc − . But dark matter does not interact electromagnetically. Λ/8πκ. ρ S 3 = constant. (4. called dark energy. Gravitational lensing is used to map the distribution of dark matter. galaxies could not have formed without it. This matter does not interact electromagnetically. In one system.Fig.7ρc . dominates. a quasar is lensed by a foreground cluster of galaxies. Exercise 4.

middle: ρ/ρc = . The points (z. Exercise 4. The function also depends on the current value of ρ/ρc . The luminosity-redshift relation. the expansion speeds up! This is due to the dark energy Λ. ) for supernovae in the range z ≈ . Then the graph becomes concave up. Fig. The graph of S(t) begins concave down.21) The graph of z vs.3.7. Exercise 4.e. Upper: ρ/ρc = 1. expresses the observed luminosity of an object as a function of its intrinsic luminosity and redshift.5 lie closest to the ρ/ρc = . 4.5 supernovae are ∼ 30% less than they would be if Λ were 0.20) implies that in the future the universe will undergo exponential expansion. This convention eliminates a constant of integration in Eq. (4. This is strong evidence that Λ = 0 . it is convenient to set t = 0 when S = 0. 2 (4. becoming dominant. 4. 77 .20).6. (4. The eﬀect of Λ is dramatic: the luminosities of z ≈ .7 billion years..4. i. lower: ρ/ρc = 0.. The best evidence comes from observations of distant supernovae. the expansion slows down. 4.5 to Fig.4.5 is the graph of for a ﬁxed L and three values of ρ/ρc . (4. i.4 Our Universe II Since the big bang is the origin of the universe. Show that Eq. Using Eqs.5. which acts as a repulsive force.5: Luminosity-redshift relaz ≈ 1. Gravity is repulsive on cosmological scales today! Exercise 4. The graph of S(t) is given in Fig. derived in Appendix 15.e.5. Show that Eq.20) implies that the age of the universe is t0 ≈ 13.12) and (4. (4. All type Ia supernovae have about the same peak intrinsic luminosity L. which dominates at early times. tion. We are now in a position to discuss the observational evidence for dark energy. 4.2. and deﬁnitely not the Λ = 0 curve. This is due to the gravitational attraction of matter.20) we can relate the redshift z of a galaxy to the time t it emitted the light we see from it: −3 (z + 1) 8πκρ0 = sinh2 Λ √ 3Λ t . t was given in Fig. The transition from the expansion slowing down to speeding up occurred at z ≈ .3 curve. The value ρ/ρc = 1 corresponds to Λ = 0 . Show that the transition from concave down to concave up occurred ∼ 5 billion years ago when the universe was ∼ 2/3 of its present size.

Summary. and 70% dark energy. and can be seen by.22) Substituting for S(t) from Eq. Each of these numbers is supported by independent lines of evidence. Consider light emitted at the big bang at event (t = 0. the light suﬀers a relative redshift z < 0 . (4. It consists of 5% known matter.9).7 billion years ago. 3 (4. the distance to rh today is dh = S(t0 ) rh . But in a universe accelerating due to dark energy. As CBR enters the gravitational ﬁeld of a cluster it gains energy.11) along the light worldline to obtain t0 dh = S(t0 ) 0 dt . The construction of a cosmological model which ﬁts the observations so well is a magniﬁcent achievement.16) for S . (4. (4. According to the discussion following Exercise 3. we may. Integrate Eq. Today dh is about 3/4 of its ultimate value. without aﬀecting the motion of a particle on the surface of the ball. the right side of Eq. The universe had an origin 13.23) Consider a ball of inertial mather of density ρ centered at the origin. more galaxies with passing time. and so we can see. Eqs. (4. As t0 increases. Our horizon consists of points at r = rh . (1) ignore the matter outside 78 .15) and (4. r = 0). They constitute 95% of the “stuﬀ” of the universe! We are in “a golden age of cosmology” in which new cosmological data are pouring in. Using Eq. From Planck’s law E = hf and the redshift formula Eq. Set Λ = 0 in the ﬁeld equations. the ﬁeld of the cluster weakens during the light’s transit through it and the loss is not quite as large as the gain. and use Eq. When leaving a static ﬁeld the radiation loses exactly the energy gained. Galaxies beyond the horizon cannot be seen by us because their light has not had time to reach us since the big bang! Conversely. in our model. with dσ = dr for k = 0. However. expanding.22) converges as t0 → ∞ and so there are galaxies which will remain beyond the horizon forever. just as a material object would. (4. We close this section by tying up a loose end from Chapter 2: We show that A = 1 in Eq. solve Eq.3 (which is based on the Schwartzschild metric and therefore only the vacuum ﬁeld equation). (4. even radical changes. Loose End. (4. (2.20) gives dh ≈ 47 billion light years. galaxies beyond the horizon cannot see us.7).16). S(t) (4. r = rh ) and received by us today at event (t = t0 . And we are left with two mysteries: the nature of dark matter and dark energy. We must be prepared for changes.6).34). the ball has radius S(t) dσ = S(t) σ (σ is constant). so does dh . (4. It is ﬂat. From Eq.4. 25% dark matter. But it crystallized only at the turn of the millennium.15): S =− 4πκρS . (1. This integrated Sachs-Wolfe eﬀect was detected in 2003. The Horizon.4 Our Universe II Further evidence for dark energy comes from observations of the CBR in directions near galaxy clusters. and the expansion is speeding up.

Thus if S(t) σ is small. which is the same as Eq.4.34). But if A = 1 in Eq. then we may apply the Newtonian equation Eq.4 Our Universe II the ball. (2. (4.23) would be multiplied by A. (2. 79 .23). The same is true of Newtonian gravitation. Einstein’s and Newton’s theories give the same result in this situation. then the left side of Eq. Thus the coincidence of the two theories in this situation requires A = 1.1) to the motion of the particle: S σ = −κ 3 4 3 π (Sσ) 2 ρ (Sσ) . (4. and (2) shrink the matter inside the ball in a spherically symmetric way to a small ball at the center.

The Robertson-Walker spacetimes illustrate these points particularly clearly. and that the expansion of the universe is speeding up. Special relativity showed that space and time are not absolute and independent. 3. Finally. we would like more evidence for such a fundamental theory. must be used. The lunar laser experiment conﬁrms tiny general relativistic eﬀects on the motion of the Earth and moon. over which an experimenter has no control. but curved and dynamic. complex. The perihelion advance of the binary pulsar and the apparent emission of gravitational radiation from its system provide quantitative conﬁrmation of the full ﬁeld equation. thought inviolate for over twenty centuries. Also. Newton’s theory is already very accurate in most circumstances. The diﬃculty in ﬁnding tests for general relativity is that the gravitational interaction is weak (about 10−40 as strong as the electromagnetic interaction). more tests possible. and radar echo delay – all conﬁrm the predictions of general relativity with an impressive accuracy. 3. the application of general relativity on the grandest possible scale. General relativity has passed every test to which it has been put. general relativity was for half a century after its discovery in 1915 out of the mainstream of physical theory because its results were needed only for the explanation of the few minute eﬀects in the solar system described in Sec. in the absence of signiﬁcant gravitational ﬁelds. gravitational lenses. They have enormously widened our “cosmic consciousness”. 80 . and so astronomical sized masses. the emission of gravitational radiation. does not apply (exactly) in a gravitational ﬁeld. and interesting universe. the CBR. light deﬂection.4. an abundantly veriﬁed theory. but also aﬀected by matter (according to the ﬁeld equation). 2. But since 1960 astronomers have made several spectacular discoveries: quasars. General relativity reduces to special relativity.2 cited experimental evidence and logical coherence as reasons for accepting the postulates of the theory. in weak gravity in which the velocity of matter is small. In consequence. Rapidly advancing technology has made. These discoveries have revealed a wonderfully varied. an exceedingly accurate theory. And we need general relativity to help us understand them. black holes. The theory explains the minute details of the motion of matter and light in the solar system. the tests of the vacuum ﬁeld equation described in Sec. not only aﬀecting matter (according to the geodesic postulate). Gravitomagnetism has probably been detected. appears to be successful. to the universe as a whole. and will continue to make. detailed structure in the CBR. Despite its fundamental nature. neutron stars. Euclidean geometry. Newton’s theory of gravity accounted for all other gravitational phenomena.5 General Relativity Today General relativity extended the revolution in our ideas about space and time begun by special relativity. As satisfying as all this is.3 – the perihelion advance of Mercury. General relativity showed that a spacetime containing matter is not ﬂat and static. but related parts of a whole: spacetime.3. Sec.5 General Relativity Today 4. This would give us more conﬁdence in applying it. the binary pulsar. General relativity reduces to Newton’s theory of gravity.

5 General Relativity Today Foremost among these are attempts to detect gravitational radiation. leptons. formidable.5 described observations of the binary pulsar which show that gravitational radiation exists. The theory of big bang nucleosynthesis uses the standard model. in 1974 Steven Hawking used quantum theoretical arguments to show that a black hole emits subatomic particles and light. and several are planned in space. time. except gravity. Despite the limited experimental evidence for general relativity and the gulf between it and quantum theory. The hope is that this radiation will someday complement electromagnetic radiation in providing information about the universe. and the forces between them. A black hole is not black! The eﬀect is negligible for a black hole with the mass of the Sun. inﬂation. The theory. We still await the ﬁrst such detection.4. dark matter. they have not been uniﬁed. String theory and loop quantum gravity are candidates for a theory of quantum gravity. and important tool in our quest for a deeper understanding of the universe in which we live. Protons and neutrons consist of three quarks. but it would be important for very small black holes. However. Despite decades of intense eﬀort. Electrons and neutrinos are leptons. Quantum theory and general relativity are the two fundamental theories of contemporary physics. and dark energy are all beyond the scope of the theory. Quantum theory describes this structure. The standard model is a quantum theory of subatomic particles and forces. Photons carry the electromagnetic force. For example. But they are separate theories. There are several kinds of each. The standard model has been spectacularly successful. Sec. For example. This and the striking beauty and simplicity – both conceptually and mathematically – of the uniﬁcation of space. and gravity in the theory make general relativity one of the ﬁnest creations of the human mind. to try to detect it directly. the goal has not been reached. correctly predicting the results of all particle accelerator experiments (although many expect discrepancies to appear soon in new experiments). forged in the 1960’s and 1970’s. it is a necessary. 81 . describes all known forms of matter. The uniﬁcation of quantum theory and general relativity is the most important goal of theoretical physics today. and force carriers. but they are very much works in progress. in cosmology the standard model leaves many matters unexplained. 3. The fundamental particles of the theory are quarks. The most unsatisfactory feature of general relativity is its conception of matter. The density ρ in the energy-momentum tensor T represents a continuous distribution of matter. entirely ignoring its atomic and subatomic structure. Several ground based detectors are in operation. Thus they must be combined on an ad hoc basis when both gravity and quantum eﬀects are important.

47 × 10−39 sec/g = 6.000 mi Mean radius Mercury orbit = 1.000.1 Physical Constants.32 sec = 6.2 Approximations.38 × 108 cm = 4000 mi Angular momentum Earth = 1020 g sec = 1041 g cm2 /sec H0 = 2.00 × 1010 cm/sec = 186.000.97 × 1027 g Radius Earth = 2.13 × 10−2 sec = 6.99 × 10 33 2 2 2 (Acceleration Earth gravity) g Radius Sun = 2.9 × 102 sec = 5.53 × 102 sec = 4.000 mi/sec (Light speed) κ = 2. The approximations are valid for small x.8 × 1012 cm = 36. (1 + x)n ≈ 1 + nx sin(x) ≈ x − x3 /6 cos(x) ≈ 1 e.2 ft/sec Mass Sun = 1.Appendices A. When appropriate.27 × 10−8 /sec = 981cm/sec = 32.60 × 106 sec = 88.3 × 10−18 /sec = (22 km/sec)/106 light year (Hubble constant) A.67 × 10−8 cm3 /(g sec ) (Newtonian gravitational constant) g = 3.96 × 1010 cm = 432.0 day Mean radius Earth orbit = 5 × 102 sec = 1.206 Period Mercury = 7. the ﬁrst number given uses (light) seconds as the unit of distance.g.. (1 + x) 2 ≈ 1 + x/2 and (1 + x)−1 ≈ 1 − x 1 82 .000 mi Mercury perihelion distance = 1. c = 1 = 3.59 × 1012 cm Eccentricity Mercury orbit = .5 × 1013 cm = 93.000 mi Mass Earth = 5.

The frequencies diﬀered by no more than one part in 1012 . longer path. See Fig.6. Thus any fractional diﬀerence between t1 and t2 would be accompanied by an equal fractional diﬀerence between f1 and f2 . A. Then t1 f1 = t2 f2 .3 The Macek-Davis Experiment. The experiment was performed to detect this eﬀect. setting up a standing wave. A fraction of the light in both directions was extracted by means of a half silvered mirror and their frequencies compared. having to “catch up” to the moving mirrors.) 83 .A.. then the clockwise beam will have a speed of light in opposite directions. In this experiment.3 The Macek-Davis Experiment A. say.6: Comparing the round trip wise. A. (If the apparatus is rotating. Let t1 and t2 be the times it takes light to go around in the two directions and f1 and f2 be the corresponding frequencies. clockFig. light was reﬂected in both directions around a ring laser. as both are equal to the number of wavelengths in the standing wave. and the frequencies of the beams will diﬀer.

but not in the parallel case. In I. Consider identical rods R and R . Thus the 84 . A. Fig. parisons to agree. R and R have the same length. First. Emit a pulse of light from A toward B . as measured in the inertial frame. According to the inertial frame postulate. A. the same time in I . then its length. B will have passed B . then its length. so R is shorter than R in I . A. each frame using the synchronized clocks at rest in that frame. B passed below B. both inertial and in motion with respect to each other. contracts. Let v be the speed of R in I. the relative speed of the light and B in I is 1 − v. situate the rods as in Fig. by observing the relative positions of the ends B and B when the rods cross. Then we can compare the lengths of the rods directly. say. then this would violate the principle: a rod moving in I is shorter. According to the principle of relativity. Since c = 1 in I. Thus in I the light will take time D/(1 − v) to reach A . A. But “at the same time” is diﬀerent to its direc. Let the length of R in I and I be D and D . For if.in I and I ! What happens is this: At the same time in I. Suppose also that the ends A and A coincide when the rods cross. There is no violation of the principle of relativity here: in each frame the moving rod is shorter. This is length contraction. Parallel case. so R is shorter than R in I.8. And at in I. say I and I . This is diﬀerent from Fig. Consider the relative positions of B Fig. respectively.7. Why this diﬀerence? An answer can be given in terms of Einstein’s principle of relativity: identical experiments performed in diﬀerent inertial frames give identical results. We show that the principle demands that the length be unchanged in the perpendicular case. while an identical rod moving in I is longer.4 Moving Rods A. B tion of motion will not have reached B.A. is unchanged. Then there can be no disagreement over the relative lengths of the rods.4 Moving Rods.7.8: Rod parallel to its direction of Thus there is no reason for the two com.motion in I. if the pulse is reﬂected at B back to A . If the rod is pointing parallel to its direction of motion in the inertial frame. where the lengths are compared directly. Perpendicular case. There is no contradiction here: The lengths are compared diﬀerently in the two inertial frames. We calculate the length contraction of R in I. Similarly. B is at distance D from A when the light is emitted.7: Rod perpendicular and B at the same time. it will take time D/(1 + v) to reach A . Now situate the rods as in Fig. as measured in the inertial frame. R and R are at rest in inertial frames. independently of any inertial frame. If an inertial rod is pointing perpendicular to its direction of motion in an inertial frame. A. Wait until A and A coincide.

(1. 1 85 .A. is contracted by a factor (1 − v 2 ) 2 of its rest length. the length of R . Since c = 1 in I and since the light travels a distance 2D in I . 1 as measured in I. + = 1−v 1+v 1 − v2 According to Eq. a clock at A will measure a total time ∆s = 1 − v 2 1 2 ∆x0 for the pulse to leave and return.12).4 Moving Rods round trip time in I is ∆x0 = D D 2D . Combine the last three equations to give D = (1 − v 2 ) 2 D . ∆s is also given by ∆s = 2D .

Set c2 − c1 = ∆c and use c1 ≈ c2 = c (say) to give N − N = 2f (D1 + D2 ) 1 1 − c1 c2 ≈ 2f D1 + D2 ∆c . Then by the deﬁnition of two way speed. f = 6 × 1014 /sec.5 The Michelson-Morley Experiment A. Rotating the interferometer 90◦ switches c1 and c2 .1) cycles out of phase. if N is a whole number + 1 . Michelson and Morley used a Michelson interferometer that splits a beam of monochromatic light in perpendicular directions by means of a half-silvered mirror. If N is a whole number. where they reunite and proceed to an observer. 2T = 2D/c. A. then the light in the two arms will reunite N =f 2D2 2D1 − c1 c2 (A. ∆c < 6 × 10−12 . c c (A. If f is the frequency of the light. D the length of the arm. light and dark will alternate N − N times. Let 2T be the time for the light to traverse an arm of the interferometer and return. The diﬀerence in the times for the two arms is 2D1 /c1 − 2D2 /c2 . giving a new phase difference 2D1 2D2 N =f − . As the interferometer rotates. the uniting beams will constructively interfere and the observer will see light. the uniting beams will destructively 2 interfere and the observer will not see light. A.5 The Michelson-Morley Experiment.9. D1 = D2 = 2100 cm. and |N − N | < 10 . The beams reﬂect oﬀ mirrors and return to the half-silvered mirror. and c the two way speed of light in the arm.2) In Joos’ experiment.9: A Michelson interferometer. c −3 86 . (A. From Eq. See Fig. c2 c1 Fig.A.2).

(A. A. the output of a laser was reﬂected between two mirrors.3) But 2T = 2D/c. Thus any fractional change in c as the slab rotates would be accompanied by an equal fractional change in f . Substitute this Fig.A.3) to give Df = cN . The fractional change in f was no more than four parts in 1015 . Fig. Then T f = N. into Eq. setting up a standing wave. A.6 The Brillit-Hall Experiment A. and N be the number of wavelengths in the standing wave.6 The Brillit-Hall Experiment. The frequency was monitored by diverting a portion of the light oﬀ the slab and comparing it with the output of a reference laser which did not rotate. The whole number N is held constant by the servo. f be the frequency of the light. Let 2T be the time for light to travel from one of the mirrors to the other and back. (A. The frequency of the laser was servostabilized to maintain the standing wave. 87 . In this experiment. This part of the experiment was placed on a granite slab that was rotated.10 shows a schematic diagram of the experiment.10: The Brillit-Hall experiment. where D is the distance between the mirrors and c is the two way speed of light between the mirrors.

and |dN | < 3 × 10−3 . according to the result of the Michelson-Morley experiment.1). to a change in the Earth’s position or speed) would result in a change dN in N : if we set.A. presumably. D1 − D2 = 16 cm.7 The Kennedy-Thorndike Experiment A. Instead of rotating the interferometer to observe changes in N in they observed the interference of the uniting beams over the course of several months. we obtain dN = 2f D1 − D2 dc c c In the experiment. (A.7 The Kennedy-Thorndike Experiment. Thus |dc/c| < 6 × 10−9 . Kennedy and Thorndike used a Michelson interferometer with arms of unequal length. 88 . f = 6 × 1014 /sec. c1 = c2 = c in Eq. Any change dc in the two way speed of light (due.

(A.5) d 2u κM +u= 2 .) Set u = 1/r.e.8 Newtonian Orbits.4) is −κM u2 . (3. A second diﬀerentiation gives the acceleration a= d 2r −r dt2 dθ dt 2 ur + r−1 d r2 dθ/dt dt uθ . a constant. dθ2 A The solution of this diﬀerential equation is u= κM [1 + e cos (θ − θp )] .. Thus −κM u2 = or d 2r −r dt2 dθ dt 2 − A2 u2 d 2u − u−1 Au2 dθ2 2 = −A2 u2 d 2u +u dθ2 (A. the coeﬃcient of ur in Eq. dt dt dθ dt dt dt Fig. dθ dθ dr dt From Eq. (Cf. (A. the coeﬃcient of uθ in Eq.20). A. (3. Eq. conservation of angular momentum. i. Cf. Eq. (2. Then d 2r d = 2 dt dθ d = dθ dθ d dr du dθ dθ = dt dθ du dθ dt dt du 2 d 2u −u−2 Au Au2 = −A2 u2 2 . 89 . Let the Sun be at the origin and r(t) be the path of the planet. 8..A. uθ = − sin θ i + cos θ j .4) is zero. At every point r deﬁne unit vectors ur = cos θ i + sin θ j .1).) This is the polar equation of an ellipse with eccentricity e and perihelion (point of closest approach to the Sun) at θ = θp . Diﬀerentiate r = rur to give the velocity v= dr dr dur dθ dr dθ = ur + r = ur + r uθ . A2 where e and θp are constants. (A.8 Newtonian Orbits A. r2 dθ/dt = A.15). (This is Kepler’s law of areas. See Fig.e.11: The polar unit vectors.4) Since the acceleration is radial. i.

(2.19) and the symmetry of f and g.A. (2. We use the notation i yr = ∂y i ∂xu ∂ 2 yi ∂ 2 xu i . xu = . xu = .9 The Geodesic Equations. Substitute the last two m displayed equations in the deﬁnition Eq. v vw ∂xr ∂y v ∂xs ∂xr ∂y w ∂y v i Diﬀerentiate y i = yq xq and use Eq. and from the assumption Eq. the formula g−1 = a−1 f ◦−1 a−1 2. (2. (fij (P )) = f ◦ .18) of ¨ jk the geodesic equations: i i i y i + Γi y j y k = yqr xq xr + yr xr y j y k = ∂k yr xr y j y k = 0 . Let (fij ) represent the metric with respect to a local planar frame at P with coordinates (xi ).18).16) to the global form Eq. (2. Apply ∂i to Eq.17) of the Christoﬀel symbols: Γi = jk 1 2 g im (∂k gjm + ∂j gmk − ∂m gjk ) i m ◦ ◦ ◦ = yr f ◦ rs ys fuv xu xv + fuv xu xv − fuv xu xv j mk m kj j km i m ◦ = yr f ◦ rs ys fuv xu xv m kj i ◦ m = yr f ◦ rs fuv xu ys xv m kj i ◦ = yr f ◦ rs fsv xv kj i = yr xr . ¨ jk ˙ ˙ kj ˙ ˙ j ˙ ˙ k j 90 .9 The Geodesic Equations A. yrs = .19). kj Substitute for y i and Γi from above to give the global form Eq. ¨ ¨ ˙ ˙ j ˙ k˙ In terms of components. and use Eq.6).12) shows that Eq. (2. (2. (2. (2. ∂k fij (P ) = 0. From Eq. ji k j ki j kj i j k m Note that (xu ys ) = (∂xu /∂xs ) is the identity matrix. Let (y i ) be a coordinate system with metric g(y i ). t from Exercise Eq. This appendix translates the local form of the geodesic equations Eq.9) holds throughout the local planar frame.5c is i m g im = yr f ◦ rs ys . (2.16): ˙ ˙ i i i y i = yq xq + yqr xr xq = yqr xr y j xq y k . (2. (2. not just at P.9). Evaluate at P : ◦ ∂i gjk = (∂p fmn ) xp xm xn + fmn xm xn + fmn xm xn = 2fmn xm xn .

The coordinates are called geodesic coordinates.17) by (gji ) and use the fact that (gji g im ) = (δj ). Multiply Eq. to obtain gji Γkp = 2 [∂p gkj + ∂k gjp − ∂j gkp ]. Given a local planar frame with coordinates x. (2. We ﬁrst express the derivatives of the metric in terms of the Christoﬀel symbols by “inverting” the deﬁnition Eq. Add the two equations to jp 2 obtain the desired expression: ∂p gjk = gji Γi + gki Γi . kp jp Deﬁne new coordinates x by ¯ xm = xm − 1 Γm xs xt .A. sj ¯ 2 sj ¯ 2 jt ¯ j ∂x ¯ ∂ xj ¯ Substitute this into Eq. (2.10 Geodesic Coordinates A. sk ¯ sj ¯ This shows already that the x are inertial frame coordinates: ¯(P ) = f (P ) = f0 . evaluate at P . ¯ 2 st ¯ ¯ where the Christoﬀel symbols are those of the x-coordinates at P .17) of the latter in terms of the m former. Exchange j and k to obtain gki Γi = 1 [∂p gjk + ∂j gkp − ∂k gjp ]. (A.10 Geodesic Coordinates. the metric f (x) = (fmn (x)) of a local planar frame at P satisﬁes (fmn (P )) = f ◦ . Thus x is a geodesic coordinate system. Diﬀerentiate this and use Γm = Γm : it ti ∂ xm ¯ ∂xm m = − 1 Γm xs − 1 Γm xt = δj − Γm xs . and apply ¯ Eq. ¯ s (A.9): ¯ x fjk (¯) = fjk (x) − fjn (x)Γn xs − fmk (x)Γm xs + 2 nd order terms. By deﬁnition. ¯ f Diﬀerentiate the above equation with respect to xp . 1 i the identity matrix.6) to f : ∂x ¯ ∂p fjk = ∂s fjk (P ) p − fjn (P )Γn + fmk (P )Γm pk pj ∂x ¯ n m = ∂p fjk (P ) − fjn (P )Γpk + fmk (P )Γpj = 0. (2. we construct a new local planar frame with coordinates x in ¯ ¯ x ¯ which the metric fmn (¯) additionally satisﬁes ∂i fmn (P ) = 0.6) 91 .

∂Gjk (E)/∂xj = 0. just as an equation of a unit circle in cartesian coordinates (say x2 + y 2 = 1) does not uniquely determine its equation in polar coordinates (r = 1 or r(r + 1) = r + 1 or .A. ). (A. .16. Thus if we transform Gjk contravariantly to other coordinate systems. (A. (A. (2. 92 . However. note that the expression of a physical law in one coordinate system does not uniquely determine its expression in another.33) holds at E in local inertial frames at E. (2. Thus there is no loss of generality in taking Eq.33) to Eq. (2.7) Is this equation Eq. We also assumed that ∂T jk (E)/∂xj = 0 in local inertial frames at E. We assume that Eq.33) is valid in all coordinate systems: Gjk = −8πκT jk . then Eq.7). This appendix justiﬁes the passage from Eq.7) the form of the ﬁeld equation in all coordinate systems? First of all.34). the validity of Eq.7) as the form of the ﬁeld equation in all coordinate systems. From Exercise 2. According to the mathematical theorem of Lovelock. (A.11 The Form of the Field Equation. (2. In the text we assumed that the Gjk depend only on the gjk and their ﬁrst and second derivatives. the right side of the equation transforms contravariantly. (A. Gjk must be of the form given in the left side of Eq.34).11 The Form of the Field Equation A. .7) in local inertial frames implies its validity in all coordinate systems. (2. From Eq.

There is no dtdr term. r) so that dτ = I(e2µ dt − udr). Substitute for dt and then for dτ in the dτ dr term: µ ν e2µ dt2 − 2udtdr − e2ν dr2 = I −2 e−2µ dτ 2 − (2u2 e−2µ + e2ν ) dr2 ≡ e2¯ dτ 2 − e2¯ dr2 . 93 .2) to eliminate the dtdr term. We make a coordinate change in Eq. r) and then a function τ (t. Then dt = e−2µ (I −1 dτ + udr).12 Eliminating the dtdr Term A.12 Eliminating the dtdr Term.A. (3. Drop the bars and rename τ to t and this becomes e2µ dt2 − e2ν dr2 . From the theory of diﬀerential equations there is an integrating factor I(t.

13 Approximating an Integral.A. 94 .13 Approximating an Integral A. (3. Retaining only ﬁrst order terms in the small quantities m/r and m/rp we have (1 − 2m/r) (1 + 2m/r) ≈ 2 1 − 2m/r 2 1 − (1 − 2m/r) (1 + 2m/rp ) (rp /r) 1− (rp /r) 1 − 2m/rp 1 + 4m/r ≈ 2 1 − (1 − 2m/r + 2m/rp ) (rp /r) 1 + 4m/r = 2mrp 2 1 − (rp /r) 1− r (r + rp ) ≈ 1 − (rp /r) ≈ 1 − (rp /r) 2 −1 −2 2 (1 + 4m/r) 1 + −1 2 2mrp r (r + rp ) 2mrp .25) can be replaced with rp t= re 1− rp r 2 −1 1+ 2mrp 4m + r r (r + rp ) dr . (3.26). Integrate to give t= Use rp 2 re − 2 rp 1 2 2 2 re + r e − rp + 2m ln rp 1 2 +m re − r p re + r p 1 2 . re to obtain Eq. 1 + 4m/r + r (r + rp ) We now see that Eq.

Our metric is now of the form ds2 = dt2 − e2ν dr2 − r2 e2λ d Ω 2 (A. (4. say S(t).r) /eλ(t . The ratio of this time to that measured at another time t is given by eν(t.r) = S(t) eν(r) .11) . (A.r) = S(t) eλ(r) . φ.r) . By isotropy at P the ratios are equal: eν(t. then the time from P in another direction to the nearby galaxy at R must also double. ds2 = e2µ dt2 . (3. and so 0 = dt2 − 2udtdr − e2ν dr2 . as in Eq. (4. Subtract the two equations to give u = 0.9) If the universe is expanding isotropically about every galaxy. r) or else there would not be isotropy half way between r and r + dr.5).r) dr. r.3).2): ds2 = e2µ dt2 − 2udtdr − e2ν dr2 − r2 e2λ d Ω 2 .11) in the form ¯ ds2 = dt2 − S 2 (t) {e2ν dr2 + r2 d Ω 2 } = dt2 − S 2 (t) dσ 2 . In Chapter 3 we showed that the metric of a spherically symmetric spacetime can be put in the form Eq. φ. θ) to a galaxy at (t + dt. (A. Eq. 0 = dt2 − 2udt(−dr) − e2ν (−dr)2 .12) 95 (A. dt2 = e2ν dr2 . θ) and Q(r + dr. Consider light sent from a galaxy at (t. r) = ν(r) to obtain eν(t. We now demonstrate this. φ + dφ) is eλ(t. Thus by Eq.9).14 The Robertson-Walker Metric. Similarly. dr = dφ = dθ = 0 for neighboring events on the worldline of a galaxy.10) cannot depend on r. θ. r + dr. Fix t . if the time from the galaxy at P to the nearby galaxy at Q doubles from t to t . (A. the ratio of the times between P and R(r. then it seems that the spatial part of the metric could depend on t only through a common factor. (A. θ). Thus e2µ = 1.r) eλ(t. Thus the metric Eq. r) to (t + dt.A.9) becomes ds2 = dt2 − S 2 (t) e2ν dr2 + r2 e2λ d Ω 2 . But ds = dt. φ.r) . r + dr) we have ds = 0. r + dr) must arrive at (t + dt. (A.10) eν(t . eλ(t. For light sent radially outward from (t.r) . Light sent radially inward from (t. Since the coordinates are comoving.14 The RobertsonWalker Metric A.r) /eν(t . since both are the time between the events measured by a clock in the galaxy. For this second signal. θ) is eν(t. (A. (A.r) e (For example.r) = λ(t . From Eq.10) are a function only of t. This appendix derives the Robertson-Walker for an isotropic universe. (A. Then both members of Eq. Similarly. (A. The coordinate change r = reλ(r) puts Eq. say S(t).8) where the coeﬃcients do not depend on φ or θ. Thus the light travel time between P (r.) By homogeneity.8). the two sides of Eq. Set ν(t . φ.

¯ r 96 . A. Set k = 1. By homogeneity. If K = 0. (A. Integrating thus gives e−2ν = C − Kr2 . See Fig. An application of Eq. The substitution r = r/(1 + k¯2 /4) gives the Robertson-Walker metric.12: Evaluating the in¯ 1 ¯ tegration constant C.12.12). Then k indicates the sign of the cur1 vature of space. (A.14 The RobertsonWalker Metric We now determine ν in Eq. and S = S |K| 2 . A. ±1). eν(0) = 1. substitute r = r |K| 2 Fig. K < 0.23) shows that the curvature K of the half plane θ = θ0 in the dσ portion of the metric satisﬁes ν e−2ν = Kr. From Eq. K = 0.A. Dropping the bars. 0. −1 according as K > 0. Thus C = 1. Since the length of the subtended arc is the radius times the subtended angle. It only remains to determine C. we obtain ds2 = dt2 −S 2 (t) dr2 + r2 d Ω 2 1 − k r2 (k = 0. (2.12) we can label the sides of an inﬁnitesimal sector of an inﬁnitesimal circle centered at the origin in the φ = π/2 plane. K is independent of r. where C is a constant of integration.

the rate at which photons are received is diminished by the same factor. (A. Thus S /S = −(z + 1)−1 z . (1. And according to Eq. We express re in Eq. From 2 2 Eqs. Deﬁne Ω0 = ρ ρc = 0 8πκ 2 ρ0 . (A. integrating. To begin. (4.14).2) with this exact relation: = L .12). Also.15 The Luminosity-Redshift Relation A. dz dr 2 2 2 = S0 H0 (1 − Ω0 ) + (z + 1)3 Ω0 . The 1 + z factors are most easily understood using the photon description of light. (4. Λ/3H0 = 1 − Ω0 . S = S0 (z + 1)−1 . from Eq. dζ (A.6).15) gives S S 2 2 = H0 (1 − Ω0 ) + Ω0 (z + 1)3 .A. (4. (4. and substituting into Eq. The energy of a photon is E = hf (an equation discovered by Einstein).14) From Eq. 3H0 where a “0” subscript means that the quantity is to be evaluated today. Recall that the critical density ρc = 3H 2 /8πκ .13) According to Eq (4. (4. Substituting in Eq. Substituting into Eq. z = (dz/dr)(dr/dt) = (dz/dr)S −1 . 2 2 [ 4πre S0 ] (1 + z)2 (A. replace the approximate relation Eq.16) [(1 − Ω0 ) + (ζ + 1)3 Ω0 ] 97 .12).15) Separating variables. 8πκρ/3 = H0 Ω0 (z +1)3 .13) gives the luninosity-redshift relation: (z) = 4π (1 + z)2 2 LH0 z 0 1 −2 . (A.7) with k = 0 .7). the term in brackets is the area of the sphere with coordinate radius re .13) in terms of z to obtain the luminosity-redshift relation.18) and (4. (4.15 The Luminosity-Redshift Relation.17). And from Eq. According to Eq. where h is Planck’s constant.11) with k = 0 . (A. the energy of each received photon is diminished by a factor 1 + z. (1. (A.

75 cosmology. 47 coordinate singularity.. 11 Fermi normal coordinates. 45 general relativity. R. 12 curved spacetime. 63. Sir Arthur Stanley. 92 frame dragging. 97 curvature. 46 Einstein. 60 deuterium. K. 59 dust. 47 crab nebula. 59. 45. 26. 59. 55.. 25 Bridgeman. 51 critical density. 78. 64 coordinates. 27. 64 blackbody. 90 geodesic postulate. 11 form of the ﬁeld equation. 24. 14 Copernicus. 73–75 cosmological constant. 13. 34 gravitational binding energy. 87 Christoﬀel symbols. 36. 13. 31. 17 Brillit-Hall experiment. 41–43 global coordinate postulate. 61 Galilei. 67 big bang nucleosynthesis.. 47. 46 ergosphere. 11 dark energy. 31.Index accelerometer. 82 balloon analogy.. 75 Eddington. 91 geodesic equations.. 74 ﬂat spacetime. 27. 69. 11 comoving coordinates. 60 geodesic. 11 geodedic eﬀect. 37 global coordinates. 32 dipole anisotropy. 42. 76–78 dark matter. Galileo. 65 event. 29 Galle. 11. 26. 29 cosmic background radiation (CBR). 81 98 . 81 black hole. 30 Gauss. 38 Einstein tensor. 44. P. B. 69 double quasar. J. 32 gravitational delay. 75 Dicke. 94 approximations. Albert. F. 48 curvature scalar. 31 approximating an integral. Nicholas. 41 clock. 32. 46 curve.. 76 gravitational radiation. Willem. 32 Brecher. 73 Braginsky. 55. 52. 31 eliminating the dtdr term. V. 58 Einstein summation convention. 63 gravitational lens. 51. 48. 75. 50 deSitter. 54. 78 degenerate electron pressure. 70 contravariant index. 90. 48 ﬁeld equation. 67 covariant index. 26. 46. 76. 55 geodesic coordinates. 35. 44. 75.W. 30. 93 energy-momentum tensor. C. 37 global positioning system (GPS). 69.

63 quantum theory. G. 13. 67 inertial force.. 14 planar frame postulate. 65. 24. 25. 16 Doppler. 29. 54. 49. 25. 31. 81 quasar. 58 photon. 59. Mercury. 17. 22. 95 horizon. 60. 33 redshift. 34. 37 postulates. 11 postulates. 86 Minkowski. 18 proper time. 58 light second. 12. Isaac. 61 length contraction. A. 19 moving rods. 73 point. 23. 37 99 LAGEOS satellites. 35 Planck’s constant. 51. 32. 61 LeVerrier. 42 loop quantum gravity. 67. 61 gravitomagnetism. 37 inertial object. 68 inertial. 72 Hubble’s law. 24 neutron star. Claudius. 66 Rebka. G. 88 Kepler. 36 local planar frame. 72 Hubble. U. 97 physical constants. G. 13. 14. 97 Planck’s law. 82 planar frame. 51. 88 Michelson-Morley experiment. 14 inertial frame postulate. 86. 64 e periastron advance. 70 Joos. B. 84 muon. 83 manifold. 24 homogeneous universe. 63 perihelion advance. 13. 14. 17. 22. 47 isotropic universe. 24. 31. 84 proper distance.. 37 Pound. 97 lunar laser experiment. 48 . 24. 30 light deﬂection. 45 Ptolemy. local form. 19. 33 expansion. 72 gravitational. 31 inertial frame. 33. 33 principle of relativity. 60 Macek-Davis experiment. 68. 21. 37. 64. 67. 46 Riemann. 53 Hils-Hall experiment. 49.. R. 36 mass. 86 Kennedy-Thorndike 24.. 33. 38–40 Michelson interferometer. 77. Johannes. 40 lightlike separated. 31 integrated Sachs-Wolfe eﬀect.. 61 Hafele-Keating experiment. 84 Lense-Thirring eﬀect. 23. 58 metric postulate. 17. 89 Painlev` coordinates. 17. 51 Newton. 92 luminosity-redshift relation. 13. Edwin. 40 pseudosphere. V. 61. 33. 17. 13. 78 intrinsic description. global form. 29 newtonian gravitational constant. 19 light speed. 81 Lovelock’s theorem. 18. 68. 29 Kerr metric.gravitomagnetic clock eﬀect. 28 pulsar. 53 newtonian orbits. 78 Hubble’s constant. 65 experiment. Hermann. 63 Ricci tensor. 18 local inertial frame. 35.

69. 54 Sloan Digital Sky Survey (SDSS). 11. 47 terrestrial redshift experiment. Karl. 66. 50 STEP. 54 tidal accelerations. 33 timelike separated. 65 stellar evolution. 77 surface. 20.. 36 time dilation. 47 white dwarf. 18 universal time. 81 static limit.Robertson-Walker metric. 33 spacelike separated. 24. 17 tensors. 75 Snider. 73. 95 rod. 33. 32 string theory. 71. 75 worldline. J. 18 spacetime. 74. 30 vacuum ﬁeld equation.L. 15. 50 WMAP. 64 Schwartzschild. 50. 15 special relativity. 11 surface dwellers. 14 spacetime diagram. 53 Schwartzschild radius. 81 supernova. 13 synchronization. 53 Sirius. 51. 50. 83 Schwartzschild metric. 11 round trip light speed. 19 spacetime coordinates. 12 100 . 12. 14. 11 standard model.

- (ebook-pdf) - Mathematics - Differential Geometry in Physicsby Ion Armeanu
- Nonlinear Connections and Nearly Autoparallel Maps in General Relativityby Sergiu Vacaru
- Foundations of Differential Geometry Vol 1 - Kobayashi Nomizuby meridine
- Relativity an Introduction to Special and General Relativityby Soham Pal

- Geometria Diferencial - Gardner - Differential Geometry & Re
- Sharipov-Course of Differential Geometry
- The Confrontation Between General Relativity and Experiment
- On Dark Energy, Weyl Geometry and Pioneer Anomaly
- Blau - Lecture Notes on General Relativity
- gravity field theory9512024v1
- general principle of relativity
- Sharipov-Classical Electrodynamics and Theory of Relativity
- generalised complex geometry
- General Relativity _2010 Hooft
- PHY4460 Relativity
- Differential Forms in Geometric calculus
- (ebook-pdf) - Mathematics - Differential Geometry in Physics
- Nonlinear Connections and Nearly Autoparallel Maps in General Relativity
- Foundations of Differential Geometry Vol 1 - Kobayashi Nomizu
- Relativity an Introduction to Special and General Relativity
- Dark Energy and Pioneer Anomaly from Weyl Geometry
- Bachman D. Geometric approach to differential forms
- Special Relativity
- Differential Geometry
- General Relativity Thooft
- Foundations of Differential Geometry
- General Relativity
- Differential Geometry
- Complex Geometry and General Relativity
- Macdonald - Elementary General Relativity
- Elementary GR
- general relativity
- Egr
- Emergence of Space-Time From Topologically Homogeneous Causal Networks 1109.0118v1
- Elementary General Relativity v3.2--Macdonald

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd