THE GOLD MODEL OF A PULSAR

Let us now return to the finding of Antony Hewish and Jocelyn Bell. They
had found rapid pulsation and the question was, what type of object could be
small enough to serve as its source. In 1968, theoreticians had two possible
candidates, the white dwarf and the neutron star, and a number of different
theories sprang up to explain the nature of pulsars. In the early days after the
discovery of CP 1919, a few more pulsars were found, thus providing further
checks and constraints on the theories; a number of these fell by the wayside
in the usual scientific competition for the survival of the fittest. In particular,
it became clear that the white dwarf could be ruled out and the neutron star,
of. much smaller size, was the more likely source. Likewise, the cause of the
pulses was found to be not the oscillations of the star but its fast spin.
A model proposed by the Cornell astrophysicist Tommy Gold in 1968
ultimately emerged as the best buy from among these theories. And although
today we still do not have a very detailed model of pulsars, the Gold model
serves as a good starting point for any more elaborate exercise to understand
them. What may be going on in and around a neutron star can be understood
according to Gold's scenario in the following way.

A neutron star has two polar axes: a rotational axis and a magnetic axis.
The Earth also has two sets of poles, one from its rotation axis and the other
from the magnetic axis. But unlike the case of the Earth, where the two axes
are nearly aligned, in a typical neutron star the two axes may be pointing in
very different directions.
The rotating star has a swarm of electrically charged particles (the
electrons) in its atmosphere. As the star rotates so does the atmosphere,
carried along by the star's strong gravitational pull. Just as the outer parts of
a merry-go-round move much faster than its inner parts, the charged
particles in the outer parts of the atmosphere move very fast, and may
approach the speed of light. For a pulsar spinning once per second, this limit
may be reached at a distance of around 50000 km from the axis of spin.
Such fast particles are known to radiate electromagnetic waves in the
presence of magnetic fields. The radiation is highly beamed like the beam of
a rotating searchlight.
So if we happen to be located in the sweep-through area of the pulsar
beam we will get pulses of radiation each time the beam sweeps past us. The
pulse period is therefore just the period of rotation of the neutron star about
its axis.
If we follow the Gold model further, we may ask the question: what
happens to the spinning neutron star as it keeps on radiating for a long time?
Obviously, the process cannot go on forever. Indeed, as time goes on, the
spinning pulsar slows down and its pulse period increases. Thus we can
imagine that the pulsar starts off spinning very fast and, as it ages, it slows
down. A pulsar which has a pulse period of one second today may slow
down to a two-second period after, say, a million years.
 Looking at pulsars of different periods, therefore, the astronomer can tell
broadly which pulsar is old and which one has just been started off. The
magnetic field also decreases as the pulsar ages and this is a factor in the
change in the intensity and spectrum of its radiation.
However, although this picture seemed to rest on reasonably secure
dations, there were further surprises in store for pulsar observers.

Superfluid rotation

The explanation of glitches has involved some remarkable physics. The slow
rotation rate show that at least 2% of the moment of inertia is attributable,
separate component of the neutron star; the long time constant of the
exponential recoveries shows that this component is a superfluid. The two
superfluid regions which might be involved are located in the core and in the
crust. Both IT involved, but in any case there must be an explanation of the
variable Coupling between the angular momentum of the fluid and that of
the crust. The rotation of a superfluid is abnormal; it is expressed as vortices,
each of which co a quantum of angular momentum. As the rotation slows,
the area density of the vortices must reduce by an outward movement. There
is, however, an interaction between the vortices and. the lattice nuclei, so
that the outward flow is impeded by the pinning of vortices to the crystal
lattice. A glitch indicates a sudden release of vortices, transferring angular
momentum to the crust. Vortices which remain pinned to the crust take no
part in the slowdown; this effectively removes part of the rotational moment
of inertia and allows the pulsar to slow down faster.

This pinning and unpinning process accounts for the steps in rotation rate
which, on average, reverse 2% of the slowdown in rotation for the bulk of
the pulsars. But, as we have seen, the glitches in the Crab pulsar are
primarily steps in slowdown rate that are not recovered between glitches.
These steps must be due to changes in magnetic dipole moment M (or in
moment of inertia I, which seems less likely). This appears to be related to
the departure of the observed values of braking index from the theoretical
value n = 3, as observed in the youngest pulsars; as noted earlier in this
section, this also can be explained by a change in M, which must be
increasing at a rate comparable with the characteristic slowdown lifetime.
The magnetic field within the neutron fluid core of the star is also quantized;
it forms flux tubes which can interact with the rotational vortices.
 An interaction between the dipolar magnetic field and the rotational vortices
can lead to an accumulation of stress which is released at a glitch, leaving a
changed magnetic field configuration.

The expansion of the rotational vortex network can carry the magnetic flux
from the core into the crust, and increase the dipole moment. It may also
stress the surface of the crust, so that the glitch may involve crust cracking
and a readjustment of the surface distribution of the magnetic field.

Radio and optical emission from pulsars

The strong magnetic field completely dominates all physical processes

outside the neutron star. The force of the induced electrostatic field acting on
an electron at the surface of a rapidly rotating pulsar like the Crab pulsar
exceeds gravitation by a very large factor, which would be as much as 1012
if there were no conducting atmosphere. The magnetic field remains
approximately dipolar out to a radial distance rc = c jw, that is, the distance
where a co-rotating extension of the pulsar, with angular velocity w, would
have a speed equal to the velocity of light. This radial distance defines the
velocity of light cylinder. Within this cylinder is a co-rotating
magnetosphere of high-energy plasma, in which the strong magnetic field
allows charged particles to move along but not across the field lines. Field
lines originating near the poles cross the velocity of light cylinder, allowing
energetic particles to escape; these particles are then able to energize a
surrounding nebula such as the Crab Nebula.

In the closed equatorial region the high conductivity allows the induced
electric field to be cancelled by a static field, so that
£ + c ^-l (w x r) x B = 0.This corresponds to a charge density in the plasma,
where the difference in numbers of positive and negative charges. is
n- - n+ = w . B(2 pi ec)^-1. The beams of radio, optical, X-ray and gamma
ray emission which provide our only observations of the pulsars originate in
the open field line region; the surface areas where these open lines originate
are known as the polar caps. At the boundary between the closed and open
regions there may be a vacuum gap in which there is a large electric field.
This region appears to have a special significance for the high-energy
radiation, which originates in electrons and positrons streaming out through
the vacuum gap. The radio emission originates closer to the surface; its
source is distributed over the polar cap.
The strong linear polarization of both the radio and the optical emission
provides valuable clues to the geometry of the emitting regions. In a typical
radio pulse the plane of polarization swings monotonically through an S-
shape; this is interpreted as the successive observation of narrowly beamed
radiation from sources along a cut across the polar cap. The plane of
polarization is determined by the alignment of the magnetic field at the point
of origin, so that the sweep of polarization can be related to the angle
between the magnetic and rotation axes and their relation to the observer.
Lyne and Manchester showed in this way that the angles between the axes
are widely distributed; there is no evidence, however, that the inclination
angle changes during the lifetime of an individual pulsar. For those pulsars
where the axes are nearly perpendicular a pulse may be observed from both
magnetic poles, while for those in which the rotation and magnetic axes are
nearly aligned the observer must be located close to the rotation axis; in this
case the radio pulse may extend over more than half of the pulse period.

The radio pulses vary erratically in shape and amplitude from pulse to pulse;
however, the integrated profile obtained by adding some hundreds of pulses
is reproducible and characteristic of an individual pulsar. Generally, these
integrated profiles contain several distinct components, known as subpulses;
these appear to be associated with different regions of the polar cap, each of
which excites radio emission in one narrowly defined direction. If the
excitation of each region varies randomly and independently of the others,
the sum will vary from pulse to pulse, but adding many pulses will produce
an integrated pulse profile which depends only on the average emission from
each region.

In many pulsars the variations of excitation are not independent in the

different regions. For example sudden changes in the observed integrated
profile seem to correspond to a switch between different patterns of mean
excitation; this behavior is known as moding. Again in many pulsars the
variation of intensity is organized into a steady drift across the profile over a
time of several pulse periods; this pulse drifting is regarded as a lateral
movement of an area of excitation across the polar cap. In some pulsars the
track of this movement appears to be closed, so that the same pattern ()f
excitation can recur after an interval considerably longer than the time for a
sub-pulse to cross the width of the pulse profile. This has been interpreted by
Deshpande and Rankin as a pattern of excitation rotating round the polar
cap.
The width of the integrated profiles varies with radio frequency. Referring
to, we note that the divergence of the polar field lines offers a natural
explanation; the lower frequencies are emitted higher in the magnetosphere
where the field lines diverge through a wider angle. A simple geometrical
model then leads to an estimated height of emission for a long-period pulsar
of about 300(v~1Hz/300)-1/2 km. Following the field lines down to the
surface, the region in which the excitation originates must only be about 250
meters across.

Since the radius of the velocity of light cylinder is determined by the angular
velocity, the scale of the magnetosphere must range over a ratio of 5000 to
1. Nevertheless the width of the integrated profiles, and the radio spectra
vary remarkably little over this range. The millisecond pulsars have broader
profiles, especially at lower radio frequencies; at high frequencies the only
difference between the radio characteristics of the two classes is in
luminosity, which is lower by a factor of ten in the millisecond pulsars. At
the other end of the scale, the 8.5 s pulsar has an exceptionally narrow
beamwidth of only 1 °: there is no explanation for this, and if it is common it
suggests that there may be a substantial population of very slow pulsars most
of which cannot be detected at all because of their narrow beams.

The radiation mechanism

The intensity of the radio emission shows at once that it must be coherent
and not thermal in any sense; the brightness temperature in some cases
exceeds 10^30 K. The optical and other high-energy radiation from the Crab
and Vela pulsars, in contrast, can be accounted for. As incoherent curvature
or synchrotron radiation from individual high-energy particles. streaming out
along field lines at the edge of the polar cap. Furthermore, the radio pulses
show a very high degree of polarization, which on occasion may approach
100%; this cannot be explained in terms of synchrotron or curvature
radiation. The radio emission is therefore coherent, and the association of a
particular frequency with a definite radial distance shows that this distance is
determined by a resonance in the plasma of the magnetosphere.
Melrose argues that a two-stage process is involved, in which the coherence
derives from bunching in an unstable stream of particles, and the radiation is
a resonant coupling at a critical density to a propagating mode directed along
a field line. The linear polarization of the radiation is then similar to that of
curvature radiation. The original acceleration of the particles takes place
near the surface of the polar cap, in a cascade process. In this cascade, as
suggested by Sturrock, electrons or positrons are accelerated to a high
energy and radiate gamma rays via curvature radiation; these gamma rays
then create electron and positron pairs as they encounter the strong magnetic
field, and the new particles are accelerated to continue the cascade.

Pulsars are most often and most easily observed at radio frequencies in the
range 0.4 GHz to 2 GHz. Their radio emission has a steep spectrum, with
spectral index a "'-' -2:1: 1, but for most pulsars observations at lower
frequencies are less informative because of the distortions due to
propagation effects. Some spectra have been measured over a wide range,
from 40 MHz to 80 GHz. There is often a turnover at frequencies below
about 100 MHz, but these spectral characteristics cannot be directly related
to the energy spectrum of the emitting particles, as can be done incoherent
sources of synchrotron radiations.

Trigonometric distance and proper motion

The parallax (giving distance) and the proper motion of nearby pulsars can
be measured. by interferometer. The most accurate measurements are made
by comparing the positions of the pulsar and an adjacent quasar, when only
differential measurements need be made. Assuming that the position of the
quasar is related to a fundamental frame, the pulsar's position may be found
in this way to submilliarcsecond accuracy. Continued observations over a
year or more provide measurements of distances through parallax, while
proper motion is also measurable for many pulsars. Geometrically
determined distances, and proper motions, may also be available from
accurate timing observations continued over a year or more.

Distances are obviously available for those pulsars which have optical
identifications, such as those in supernova remnants. Some distance
information is available for pulsars close to the Galactic plane which are
observed through neutral hydrogen clouds. These H I clouds absorb at
wavelengths near 21 cm; the actual absorption wavelength depends on their
location, giving an indication of the distance of an individual pulsar.
These distances are useful as calibrators of the distances provided by the
dispersion measure (DM). Given a model of the distribution of electron
density in the interstellar medium, the DM gives the distance. For nearby
pulsars an average electron density ne = 0.025 cm-3 pc may be used, with an
accuracy of around 20%; for example the Crab pulsar with DM = 57 is
known to be at a distance of about 2 kpc. At distances greater than 1 kpc it is
however necessary to take account of the structure of the Galaxy; a model
distribution by Taylor and Cordes is generally used for such distance
determinations.

X-ray pulsars

The explanation of the new class of binary and millisecond pulsars, and their
linkage to the main population came from X-ray astronomy. The strongest
X-ray sources are binary systems in which a neutron star is accreting matter
from a companion star; in the process the neutron star is spun up and can
eventually become a millisecond pulsar. These are thermal sources of X-
rays; the accreting gas is concentrated by the magnetic field, forming hot
spots above the magnetic poles. The X-ray flux then varies as the star
rotates, bringing the hot spots in and out of view. (A different class of X-ray
pulsars, such as the Crab pulsar, is radio pulsars whose spectrum of pulsed
radiation extends continuously from radio up to optical, X-ray, and even
gamma-ray energies.)

The identification of an X-ray source with a condensed star was first demon-
strated by the discovery of pulsations at intervals of 4.8 s from the source
Cen X-3; furthermore, a periodic Doppler shift in the pulse interval showed
that the source was in a binary system. More than 100 such X-ray sources
are now known, distributed throughout the Galaxy. All are neutron stars in
binary systems; their companions have masses between 1 and 20 MG. Most
of the companions are young stars with masses between 5 and 20 MG,
which are evolving through the red giant phase towards a supernova
collapse, and whose remnant core would probably become another neutron
star. Others, the low-mass X-ray binaries (LMXBs), have companions with
masses about 1 MG, which are evolving towards a collapse to a white dwarf.
Twelve of the 50 LMXBs are located in globular clusters, indicating a link
with the binary and millisecond pulsars.
The binary radio pulsars are similarly revealed by the periodic Doppler shift
of their pulse periods. 12.17 shows this for two examples with orbits of
small and high eccentricity. All of those with small eccentricities have a
low-mass white dwarf companion; these binaries are the likely outcome of
evolution from LMXBs. Those with high eccentricity usually have a neutron
star companion; these appear to have evolved from the higher-mass systems,
the HMXBs. Solitary millisecond pulsars may have evolved from either
class; in the case of the HMXBs the binary may have disrupted at the time of
the supernova explosion, while for the LMXBs the white dwarf may have
been evaporated by intense radiation from the pulsar itself. The latter
scenario is supported by observations of several pulsars with very-low-mass
companions, in which an occultation occurs over a large part of the binary
orbit; this is attributed to a cloud of ionized gas streaming away from the
white dwarf.

The increased rotation rate of the millisecond pulsars is driven by transfer of

angular momentum from the orbit during the mass transfer. There seems to
be a lower limit of 1.5 ms on the period, although this may still be an effect
of observational selection. If this limit becomes established, it may be
related either to the influence of the neutron star magnetic fields on the
accretion process, or to the stability of the neutron star itself. A 'hard' rather
than 'soft' equation of state leads to a larger diameter, with a larger disruptive
centrifugal force. If searches confirm the present limit on rotation rate, this
will support the hard equation of state.

Magnetic dipole moments

The magnetic strength of a pulsar is expressed as the polar field strength Bo

of an orthogonal dipole, calculated from the slowdown rate; for example, the
small slowdown of the millisecond pulsars indicates that their magnetic
dipole moments are smaller than for the normal pulsars. 12.18 shows the
distributi9n of polar field strengths in all pulsars, calculated from measured
values of P and P. The youngest pulsars are those associated with supernova
remnants; these have large magnetic fields. 12.18 also includes some X-ray
pulsating sources known as the magnetars; these also are neutron stars, with
the highest magnetic
dipole fields but comparatively long periods. Several are associated with
supernova remnants. There may be an association between the magnetars
and several pulsars, which appear close to them in the diagram, but no
evolutionary pattern has yet emerged. The magnetars are distinguished by
their very high X-ray luminosity, which demands a larger source than the
decay of rotational energy. It is proposed that the source of energy for the
observed X-rays may be from the decay of the large magnetic field rather
than from rotational energy.

The normal evolution of solitary pulsars in this diagram, according to the

slowdown law without any change in the magnetic field, would follow a
horizontal track to the right, eventually reaching the death line as the rotation
rate fell below a critical value. The dipole moment of the youngest pulsars
may be increasing, but the actual distribution of normal pulsars in this
diagram suggests instead that there is field decay in the normal population.
The millisecond pulsars, which are much older, have much lower fields;
there may be some process related to the binary spin-up, which reduces the
field from the order of 10^12 gauss to about 3 x 10^8 gauss.

PULSARS AS STANDARD CLOCKS

We have already remarked on the fact that the time period of the first pulsar,
CP 1919, could be quoted to ten decimal places. The unusually steady
periods of pulsars, especially the millisecond pulsars which were discovered
in the 1980s, opened up the possibility that pulsars could serve as the basic
time keepers for natural phenomena.
The present definition of Universal Time (UT) is in terms of the idealized
caesium clock. This clock depends on the oscillations of the caesium atom.
In practice the second is defined as the duration of 9192631770 periods of
the radiation corresponding to the transition between two specified states of
the caesium atom. The characteristic time intervals associated with each
such atomic transition are, however, not strictly the same. But, by averaging
over several such clocks, one can arrive at a steady time period. Pulsars,
however, seem to do better in terms of providing us with a steady time
standard, as can be seen below.
The quantitative estimate of the steadiness of a clock is provided by the
so-called Allan variance of its errors. To obtain this variance, one measures
the fluctuations in time period as a fraction of the latter and averages the
squares of such fluctuations. This variance decreases if we can measure it
over a longer interval of time, provided we are confident that the basic time
period remains stable over the interval of measurement. Thus the longer this
time interval, the less is the Allan variance and the male accurate is the
clock.
For a caesium clock the interval is of the order of a month. In 4.16 we see
how the variance falls over a period of a million seconds or so and then
begins to rise. By contrast, the same shows that for the pulsar PSR 1937+21
the time interval of steady period runs into years! That is, over short time
scales of a month or so the pulsar may not do as well as the atomic clock.
But for longer durations its stability tends to overtake the latter, making it
accurate to thirteen decimal places.

Thus one could proceed as follows to construct a purely pulsar-based time

standard. Suppose one could show, by making observations of the smoothest
running pulsars, that the difference between the time standards given by two
pulsars is less than the difference between UT and the mean pulsar standard.
If so, one could rely entirely on pulsars as the basic timekeepers. This would
certainly improve the UT standard by bringing down the fluctuations within
which it is currently measured. Whether pulsars will ultimately replace
atomic clocks is an open question.

PULSARS AND THE TESTS OF GRAVITY THEORIES

The very accurate time keeping by pulsars has helped physicists in

another way. The binary pulsar PSR 1913+16 introduced earlier has proved
to be extremely useful in testing the predictions of Einstein's general theory
of relativity vis-à-vis some other gravity theories. .
The general theory of relativity has a starting point very different from the
much simpler Newtonian law of gravitation but for most practical purposes
it ends up giving the same answers. Thus, to test which of the two theories is
closer to reality one needs more subtle tests that require very precise
measurements and rather special circumstances. Such tests, within our own
solar system, have been primarily responsible for increasing the credibility
of general relativity at the expense of Newtonian gravitation. These tests,
however, require very accurate measurements.

The advance of periastron

For example, take the test provided by the motion of the planet Mercury
around the Sun. shows that according to Newtonian gravitation Mercury
should move in an elliptical orbit with the Sun as a focus of the ellipse.

However, in practice, the motion of Mercury is a little more complicated,

as shown in. The line joining the Sun to the perihelion, the closest point in
the orbit, slowly changes direction with time.
 This strange behavior had been noticed in the last century and several
attempts were made to understand it within the framework of Newtonian
gravitation. Thus a substantial fraction of this movement of the line was
known to be due to the gravitational pull on Mercury by other planets of the
solar system, especially Venus, Earth and Jupiter.

How tiny was the anomaly could be seen as follows shows the type of
protractor used in a school mathematics kit. It measures angles in tiny
divisions marked on its circular boundary. Each division is a degree. Divide
the degree into 60 equal parts to get an even tinier measure of angle called a
minute of arc. Next make 60 divisions of a minute of arc to get a second of
arc (an arcsecond). The anomalous shift of Mercury's perihelion as viewed
from the Sun was at the rate of 43 seconds of arc in 100 years.
Tiny though this discrepancy looks, it was sufficient to worry the theorists,
who up till then had found Newton's law of gravitation in complete
conformity with observation. And it was here that general relativity stepped
in with the right answer. It introduced a small modification in the way a
planet moves round the Sun and showed that it exactly accounted for the
anomalous 43 seconds of arc per century.
Apparently, disgressed from pulsars to planets, to show how tiny, yet
significant, was the difference between two gravitational theories, one of
Newton, the other of Einstein. It is against this backdrop that we have to
view the enormous improvement in time measurement provided by binary
pulsars.
The two stars in PSR 1913+I6 move in a binary system, each following an
elliptical orbit. The line joining them, however passes through a fixed point
in space called the centre of mass of the pair.7'Of course, while the centre of
mass stays constant, the distance between them varies. Like the perihelion in
the case of the Sun, we may talk of a periastron for binary stars when the
distance between them is least.

Actually the Sun-Mercury system could also be viewed as a binary system.

The mass of the Sun, however, is some six billion times the mass of
Mercury. The result of this huge ratio of masses is that the Sun hardly
budges under Mercury's tiny force of attraction: their centre of mass is
almost coincident with the centre of the Sun.
This particular circumstance enables the relativist to calculate the rate of
advance of the perihelion of Mercury almost exactly. In the case of a binary
pulsar, the situation is different. The two stars (pulsar A and companion B)
are of comparable mass and so a repeat of the calculation in exact terms is
not possible. The so called two-body problem, in which two comparable
masses move under each other's gravitational pull, is extremely difficult and
has not been solved in general relativity.
Nevertheless, one can make approximate calculations which are
considered credible by pundits in the field. These calculations give a value
for the advance of periastron of PSR 1913+16 of the right magnitude, 4.2
degrees per year, as observed. (Notice that this effect is about 350000 times
that observed for Mercury.) The binary pulsar therefore provides a
confirmation of general relativity through the observed shift of periastron.

Time delay
Another effect peculiar to general relativity (and not found in Newtonian
gravitation) is to do with the time delay in a light signal passing close to a
massive body. We will see in the next chapter how general relativity requires
space-time measurements to be modified in the neighbourhood of such a
body, because of its gravitational influence. Thus the to-and-fro passage of a
radar signal would have a longer duration if such modifications were
present.
In the solar system this effect was observed by the Mariner spacecraft in
bouncing radio signals off the surface of Mars, when these signals grazed the
Sun. Compared to the situation when the Sun was nowhere near such
signals, the delay was about 250 microseconds.
In the case of the binary pulsar PSR 1913+I6, the pulsar signal will take
about 50 microseconds longer to reach us when it grazes. the pulsar's
companion. The effect, though small, can be accurately measured, thanks to
the precise time keeping of the pulsar. And the measurements have
confirmed the above prediction of general relativity.

Existence of gravitational radiation

Although these tests have been very fruitful in pushing up the credibility of
general relativity one notch higher, none of them has generated as much
excitement as the reported (indirect) evidence for the existence of
gravitational waves. Let us first look at how these waves are expected to be
produced.
An analogy with electromagnetic waves will help. The most basic
mechanism for the emission of such waves is an oscillating electric charge.
The to-and-fro motion of such a charge will generate energy in the form of
electromagnetic waves. A detector of electromagnetic waves can easily
measure this radiation. However, we can indirectly infer its existence by
asking the question: Where is this energy coming from? It has to come from
the motion of the electric charge. Therefore, as the charge continues to
radiate, its motion slows down, much like a car slowing down against
friction of the road when its engine is turned off. So, from the damping of its
motion, we can deduce that the electric charge has been radiating energy.
Just as an oscillating charge radiates electromagnetic waves and slows
down as a result, so do we have massive dynamical systems radiating
gravitational waves and slowing down. In theory, that is, for no one has yet
succeeded in directly detecting gravitational waves.
 General relativity tells us that the simplest system radiating gravitational
waves would be a binary system, where two masses go round each other, as
in the case of the binary pulsar PSR r9r3+I6. It is expected that because of
the energy lost in gravitational radiation, the binary pulsar system should
shrink. That is, the two members should move around each other in smaller
and smaller orbits. As the orbits shrink, the binary period decreases.
The theoretically estimated decrease is as small as 2.4 Pico seconds per
second! (One million million Pico seconds make one second.)
Nevertheless, thanks to the accurate timing kept by the pulsar, this minute
effect has been measured and verified. One can see this by observing the
cumulative change of phase in the orbit, which was two seconds over six
years. 4.23 shows a graph of such observations.
The fact that the orbital period has decreased is apparent from this phase
change and the steady rate of decease is seen as a confirmation of the
general relativistic prediction that such binaries should radiate gravitational
waves. There. are other theories making similar predictions, although of
different amounts, but the accuracy of measurement of this binary system is
such that it rules out such alternatives.
We may have to wait until early in the twenty-first century for a direct
proof that gravitational waves exist. Several large detectors are currently
under construction with the specific aim of picking up gravitational waves
by binary stars in orbits that progressively shrink until the stars coalesce.
The binary pulsar, however, reassures us that such waves do exist!

PLANETS AROUND PULSARS

Normally, therefore, one expects to find planets around stars like the Sun,
which have been steadily producing energy through the fusion of hydrogen
to helium. Indeed a few cases of such stars with planets are now known.

In 1991, however, there was a claim that a planet had been found around a
pulsar! Keeping in view the somewhat traumatic origin of a pulsar, how can
it manage to acquire a planet? Surely, any planets that a star may have had
before it became a supernova would have been blown away or destroyed by
the explosion. Therefore, when, in 1991, a group of radio astronomers from
Jodrell Bank announced that a particular pulsar did seem to have a planet
around it, the news came as a complete surprise.
How did the astronomers make this discovery? The pulsar signals seemed
to show a small wobble that could only be explained if the pulsar had a
planet around it, gravitationally disturbing it. The situation is somewhat
similar to that for binary stars, where each star affects the motion of its
companion; but, in this case, the planet being far less massive than the star
manages to produce only a barely perceptible effect. Thus the star will
wobble slightly as the planet orbits it. The extent and period of the wobble,
if measurable, can tell us something about how massive the planet might be
and how long it takes to orbit the star. (Remember, the planet itself, being
non-luminous, is not visible.) Thus the Jodrell Bank group was relying on
this indirect evidence for making their claim.
The announcement of the discovery was, of course, an immediate sensation.
As is not uncommon when such unexpected findings are announced, a
special conference was later arranged to discuss the detail and implications
of this discovery. But the discovery itself turned out to be a false alarm!
Suspicion as to its validity in fact arose when it was discovered that the
predicted planet seemed to have a period of six months or a year, exactly
matching the Earth's period! Ultimately it turned out that because we are
observing the pulsar from the moving Earth, our motion also affects the data
and produces the periodic pattern. So, this was no real effect: it was simply
the result of observing a pulsar from a moving platform. Ironically, at the
conference at which this discovery was retracted by Andrew Lyne from
Jodrell Bank, Aleksander Wolszczan, an astronomer working at the radio
telescope at Arecibo in Puerto Rico, reported that he had found a pulsar
with two planets. The pulsar had the catalogue number PSR 1257+12.
Now, if you have had. A false alarm once you are less inclined to believe a
similar case. You will obviously ask for a check and double-check of the
records. But W olszczan had already taken sufficient care to see that the
effect of the Earth's motion was accounted for and that any other spurious
effect was removed. Thus he was confident of the reality of the effect, which
was also double-checked by others.
Thus, at least two planets are known to move around this particular pulsar;
one has a mass of 2.8 times the mass of the Earth and the other a mass of 3.4
times the mass of the Earth. Their respective periods around the pulsar are
66.6 days and 98.2 days. So they are moving relatively fast, like Venus or
Mercury. Their respective distances from their parent star are 70 million
kilometers and 54 million kilometers. That is, they are relatively close by.
(Just for comparison the Earth is orbiting the Sun at a distance of ISO
million kilometers.) Now observers have claimed that there is a third planet
also in that system. But we still do not know how those planets got there.
This is a problem for the theoreticians to worry about!