This page intentionally left blank

Beyond Pluto
Exploring the outer limits of the solar system
In the last ten years, the known solar system has more than doubled in size. For
the rst time in almost two centuries an entirely new population of planetary
objects has been found. This Kuiper Belt of minor planets beyond Neptune
has revolutionised our understanding of how the solar system was formed and
has nally explained the origin of the enigmatic outer planet Pluto. This is the
fascinating story of how theoretical physicists decided that there must be a
population of unknown bodies beyond Neptune and how a small band of
astronomers set out to nd them. What they discovered was a family of ancient
planetesimals whose orbits and physical properties were far more complicated
than anyone expected. We follow the story of this discovery, and see how
astronomers, theoretical physicists and one incredibly dedicated amateur
observer have come together to explore the frozen boundary of the solar
system.
JOHN DAVI ES is an astronomer at the Astronomy Technical Centre in
Edinburgh. His research focuses on small solar system objects. In 1983 he dis-
covered six comets with the Infrared Astronomy Satellite (IRAS) and since
then he has studied numerous comets and asteroids with ground- and space-
based telescopes. Dr Davies has written over 70 scientic papers, four astron-
omy books and numerous articles in popular science magazines such as New
Scientist, Astronomy and Sky & Telescope. Minor Planet 9064 is named
Johndavies in recognition of his contributions to solar system research.
Beyond Pluto
Exploring the outer limits
of the solar system
John Davies
iuniisuio n\ rui iiiss s\xoicari oi rui uxiviisir\ oi caxniioci
The Pitt Building, Trumpington Street, Cambridge, United Kingdom
caxniioci uxiviisir\ iiiss
The Edinburgh Building, Cambridge CB2 2RU, UK
40 West 20th Street, New York, NY 10011-4211, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
Ruiz de Alarcn 13, 28014 Madrid, Spain
Dock House, The Waterfront, Cape Town 8001, South Africa
http://www.cambridge.org
First published in printed format
ISBN 0-521-80019-6 hardback
ISBN 0-511-03214-5 eBook
Cambridge University Press 2004
2001
(Adobe Reader)

To all my friends in Hawaii, especially those who made UKIRT the

greatest infrared telescope in the world.
Contents
Preface page ix
Acknowledgements x
Prologue xii
1 The edge of the solar system 1
2 The Centaurs 21
3 The mystery of the short-period comets 39
4 Shooting in the dark 47
5 Deeper and deeper 71
6 Sorting out the dynamics 93
7 What are little planets made of ? 115
8 Numbers and sizes 139
9 Things that go bump in the dark 151
10 Dust and discs 167
11 Where do we go from here? 179
12 Will we ever get our names right? 191
Appendix 1 Dramatis personae 211
Appendix 2 Guidelines for minor planet names 225
Index 229
vii
Preface
This is a story about a discovery and some of the developments which
followed it. It is not a textbook. Although I hope it contains most of the
relevant technical details I set out to show a little of how astronomy is
actually done. Some of the characters spend their time looking
through telescopes on the darkest of dark nights, others work in
ofces and laboratories far removed, both physically and psychologi-
cally, from mountaintop observatories. From time to time this diverse
group of people come together, in small groups or en masse, to
exchange ideas and dispute data. They do this in order to understand
the origin and evolution of the solar system in which we live and
work. A few names crop up frequently, for the community of solar
system astronomers is a small one and our paths often meander
across each other in unpredictable ways.
In the last few years a new, and dynamic, outer solar system has
replaced the sterile border known to our predecessors. I still nd it
hard to believe how much our view of the solar system has changed in
the last decade and even harder to credit that I have been a part of this
adventure. It has been an exciting time for all of us, and some of my
childhood dreams have come true in a way that I could never have
imagined. I hope that some of this mystery and excitement comes
through in these pages.
ix
Acknowledgements
Although I put this book together, many of the ideas in it sprang from
the fertile minds of my fellow astronomers. It is they who have opened
up the distant frontier of the solar system. Many of them have shared
their thoughts, and in some cases their images, with me over the last
couple of years. Some consented to be interviewed in person, others
tolerated a barrage of email enquiries, and all of them seemed to have
taken it in good humour. A good fraction of them read specic sec-
tions of the draft manuscript (a few brave souls tackled the whole
thing) and put me right when I strayed from the facts. They also
helped make clear to me things that were uncertain. I enjoyed the spir-
ited debate about the planetary status of Pluto, and the sharing of rec-
ollections or thoughts about events that happened a long time ago. Id
like to thank them all for their help. In particular:
Mike A Hearn, Mark Bailey, Gary Bernstein, Alan Boss, Mike Brown,
Robert (Bob) Brown, Marc Buie, Al Cameron, Anita Cochran, Dale
Cruikshank, Cathy Delahodde, Martin Duncan, Julio Fernandez, Tom
Gehrels, Brett Gladman, Dan Green, Simon Green, Jane Greaves, Oli
Hainaut, Eleanor (Glo) Helin, James Hilton, Wayne Holland, Piet Hut,
Dave Jewitt, Charles Kowal, Jane Luu, Renu Malhotra, Brian
Marsden, Neil McBride, John McFarland, Steve Miller, Warren Offutt,
Joel Parker, Dave Rabinowitz, Eileen Ryan, Glenn Schnieder, Jim
Scotti, Carolyn Shoemaker, Ronald Steahl, Alan Stern, Steve Tegler,
Rich Terrile, Dave Tholen, Nick Thomas, Chad Trujillo, Tony Tyson
and David Woodworth.
Material in Appendix 1 is reproduced with the permission of the
Minor Planet Center and Lutz D. Schmadel. A full list of Minor Planet
x
citations may be found in the Dictionary of Minor Planet Names, by
L. D. Schmadel, which is published by Springer Verlag and is cur-
rently in its 4th edition.
Many of the illustrations came directly from the scientists
involved and are credited in the captions. Figures 3.8, 6.2, 7.2, 7.4 and
8.1 are reproduced from Protostars and Planets IV (editors Vince
Mannings, Alan Boss, Sara Russell) with permission of the University
of Arizona Press, 2000 The Arizona Board of Regents, courtesy of
Dave Jewitt. Figure 2.4 was published in Icarus and appears with per-
mission of Academic Press. I would also like to thank the following for
their help with specic images or information.
George Beekman, Zenit Magazine.
Germaine Clark & Brooke Hitchcock, Lawrence Livermore
National Laboratory.
Richard Dreiser, Yerkes Observatory.
Cheryl Grundy, Space Telescope Science Institute.
Debora Hedges & Lynne Nakauchi, California Institute of
Technology.
Bob Milkey, American Astronomical Society.
Lori Stiles, University of Arizona.
Richard Wainscoat, Institute for Astronomy, Honolulu.
Cynthia Webster, Lowell Observatory.
Remo Tilanus for translating George Beekmans article into a lan-
guage I understand and to Tim Jenness for converting those GIFs,
TIFFs, JPEGs and bitmaps into things my printer could understand.
Adam Black, of Cambridge University Press for getting me into
this, and to Alice Houston, Simon Mitton and Alison Litherland, also
of the Press, who got me out of it.
Finally, my wife Maggie for reading and correcting the manuscript,
even though its not really her subject.
A
c
k
n
o
w
l
e
d
g
e
m
e
n
t
s
xi
Prologue
In July 1943 the Journal of the British Astronomical Association pub-
lished a short article entitled The Evolution of our Planetary
System. The paper had been submitted by a retired Irish soldier and
part-time amateur theoretical astronomer, Lt-Col. Kenneth
Edgeworth. Despite being greatly reduced in length due to wartime
shortages of paper, the article contained a prophetic paragraph on the
structure of the solar system. While discussing comets, Lt-Col.
Edgeworth remarked, It may be inferred that the outer region of the
solar system, beyond the orbits of the planets, is occupied by a very
large number of comparatively small bodies. Kenneth Edgeworth did
not live to see his prediction conrmed, but almost 50 years later just
such an object was discovered. This new body, initially called simply
1992 QB
1
, was the harbinger of a breakthrough in our understanding
of the solar system. Within a few years hundreds of similar objects
would be found in what, by an ironic twist, soon became known as the
Kuiper, rather than Edgeworth, Belt.
xii
The edge of the
solar system
Like the planets Pluto and Neptune, the existence of what is fre-
quently called the Kuiper Belt was predicted theoretically long before
it was actually observed. Probably the rst fairly detailed speculation
about a cometary ring beyond Neptune was put forward by Kenneth
Essex Edgeworth in 1943. Edgeworth was an interesting character
who had progressed from soldier and engineer to retired gentleman
and amateur theoretical astronomer. He was born on 26th February
1880 in County Westmeath, Ireland, into a classic well to do literary
and scientic family of that era. As a young man he joined the Royal
Military Academy at Woolwich, England, and attained a commission
in the Royal Engineers. He spent his next few years stationed around
the world building bridges, barrack blocks and the like. With the out-
break of the First World War he served with the British Army in
France as a communications specialist and was decorated with both
the Distinguished Service Order and the Military Cross. He remained
in the army until 1926 and then took up a position with the Sudanese
department of Posts and Telegraphs in Khartoum. Edgeworth
remained in the Sudan for ve years before retiring to Ireland to live
out the remainder of his life.
Although retired, Edgeworth was by no means inactive. During the
1930s he studied economic theory and published several books on this
topic. Although never afliated with a university or other astronomi-
cal institution, he also pursued an interest in astronomy which he had
acquired in his youth (he had joined the Royal Astronomical Society
in 1903). After he retired he wrote a number of articles, mostly theo-
retical in nature, dealing with the process of star formation and devel-
oping ideas about the origin of the solar system. In 1943 he joined the
British Astronomical Association, whose journal published his rst
C
H
A
P
T
E
R
1
1
paper on the evolution of the solar system that summer. It was a short
note which Edgeworth himself described as containing Not so much
a theory, but the outline of a theory with many gaps remaining to be
lled. In his paper he described how a cloud of interstellar gas and
dust might collapse to form a disc. He speculated that within such a
disc numerous local condensations of higher density might then
develop and collapse upon themselves. Noting that the real solar
system does not comprise a huge number of small objects, but rather a
few large planets and moons, Edgeworth suggested that these conden-
sations then coalesced to form the nine known planets and their satel-
lites. Crucially, Edgeworth recognised that there was no obvious
reason why the disc of planet-forming material should have been
sharply bounded at the orbit of the outermost planet. He suspected
that the disc probably extended far beyond this distance and reasoned
that, so far from the Sun, the density of material in the disc would be
B
e
y
o
n
d

P
l
u
t
o
2
Figure 1.1 A caricature of Kenneth
Edgeworth as a young man.
Comparison with photographs of
him in later life suggests that it is a
good likeness. (Royal Signals
Museum Archive.)
very low. So, although individual condensations of reasonable size
might form beyond Neptune, there would be little likelihood of them
encountering each other frequently enough to form large planets. He
suggested instead that these condensations would simply collapse
upon themselves to form a large number of small bodies. Echoing
then current theories of comets as concentrated swarms of mete-
oroids he described these distant condensations as astronomical
heaps of gravel. He added that perhaps from time to time one of these
condensations Wanders from its own sphere and appears as an occa-
sional visitor to the inner solar system. Here was the genesis of the
idea of a trans-Neptunian comet belt.
Edgeworth developed his ideas further, writing a longer paper
along similar lines a fewyears later. This second paper was submitted
to the Monthly Notices of the Royal Astronomical Society in June 1949.
Although by todays standards it contained numerous poorly justied
assumptions, it was accepted almost immediately and appeared in an
issue of the journal dated late 1949. In this paper Edgeworth expanded
on his model for the formation of the planets and once again men-
tioned the likely existence of a vast reservoir of potential comets
beyond the orbit of Neptune.
About the same time as Edgeworths musings, the Dutch-born
astronomer Gerard Kuiper was also considering the existence of tiny
worlds beyond Pluto. Kuiper was working at the Yerkes Observatory
in Chicago and, in 1951, he wrote what became a classic book chapter
summarising the state of knowledge about the solar system. Kuiper
noted that the distribution of material in the outer solar system
seemed to come to an unnaturally sharp edge in the region of the
planet Neptune and that there was no obvious reason why this should
be so. Perhaps taking a lead fromnewly published theories about the
composition of comets, Kuiper suggested that during the formation of
the planets many thousands of kilometre-sized snowballs might
have been formed in a disc beyond the planet Neptune. Like
Edgeworth, Kuiper reasoned that at such great distances fromthe
Sun, where the relatively tiny snowballs would occupy a huge volume
of space, it was unlikely that these snowballs could come together to
formlarge planets. He suggested that instead their orbits were dis-
turbed by the gravitational inuence of the planet Pluto

and they
were either ejected into deep space or sent in towards the Sun to
T
h
e

e
d
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
3

Pluto was then thought to be much more massive than we now know it to be.
appear as comets. However, in a world in which observational astron-
omy was still dominated by the photographic plate, the detection of
such tiny objects remained impracticable.
Of course, speculation about missing planets is not a new phenom-
enon. Ever since William Herschels discovery of Uranus in 1781,
astronomers have been fascinated by the possibility that there might
be other unknown worlds. On the 1st of January 1801 the Italian
astronomer Giuseppe Piazzi made a chance discovery of what was at
rst thought to be a new planet. The object, which was soon shown to
be orbiting between Mars and Jupiter, was named Ceres after the
Roman goddess of the harvest. It was soon found that Ceres, even
though it was quite close, did not show a detectable disc when viewed
through a telescope. This suggested that it was smaller than any of the
other known planets. Three similar objects, Pallas, Juno and Vesta,
were found in 1802, 1804 and 1807 respectively. All appearing as slow-
moving points of light, this group of new objects was referred to as
asteroids (star-like) by William Herschel. All went quiet for a while
until the mid 1840s when new asteroids began to be found in quite
large numbers. By the end of 1851 fteen of them had been found and
we now know that Ceres is just the largest of many small rocky
objects in what became known as the asteroid belt.
However, by the middle of the nineteenth century attention had
once again swung to the outer solar system. Irregularities in the
motion of Uranus hinted that it was being tugged by the gravitational
pull of another, more distant world still waiting to be discovered. In a
now classic story of astronomical detective work, the mathematicians
Urbain Le Verrier of France and John Couch Adams of England inde-
pendently calculated the position of the unseen planet, making its dis-
covery a relatively simple matter once someone could be persuaded to
look in the appropriate direction. In the event, it was Le Verrier whose
prediction was rst tested. While Adams calculations lay almost
ignored by the English Astronomer Royal, the director of the Berlin
Observatory J. G. Galle and his assistants searched the region sug-
gested by Le Verrier. They found the predicted planet on 23rd
September 1846. However, the discovery of Neptune was not the end of
the issue as far as distant planets were concerned. After a few decades
it seemed that Neptune alone could not explain all the problems with
the orbit of Uranus. This hinted that there might be yet another
planet lurking in the darkness of the outer solar system. Two
Americans set out to see if this was the case.
B
e
y
o
n
d

P
l
u
t
o
4
William Pickering was one of these planet hunters, suggesting in
1908 that a planet with twice the mass of the Earth should lie in the
direction of the constellation Cancer. His prediction was ignored.
Eleven years later he revised his calculations and pointed to a position
in nearby Gemini. This time astronomers at the observatory on Mt
Wilson, California, responded, using their 24 cm telescope to search
around Pickerings predicted coordinates. They failed to nd any-
thing. Meanwhile, American millionaire Percival Lowell was also
turning his attention to the outer solar system. Lowell, who had
earlier convinced himself that intelligent life existed on the planet
Mars, rmly believed that deviations from the predicted positions of
Uranus meant that there must be another unseen planet remaining to
be discovered. He called this distant object Planet X and, like
Pickering, he tried to calculate where in the sky it might be found.
However, Lowell had an advantage over his rival, for he had the means
to pursue his search without relying on the whims of others. Lowell
owned a private observatory which he had founded in 1894. It was
built on Mars Hill, just outside the town of Flagstaff, Arizona. Unlike
modern observatories, which are usually located on barren mountain-
tops, Lowell placed his telescopes in a delightful setting. The Lowell
observatory was surrounded by pine trees and had a ne view back
across the town.
Lowells Planet X was predicted to be quite large, but very distant,
and so was unlikely to show an obvious disc in the eyepiece of a small
telescope. The best way to nd it would be to detect its daily motion
relative to the xed background of stars and galaxies. In the previous
century such searches had been made by laboriously sketching the
view through a telescope and then comparing this with sketches of
the same region made a few days earlier. However, by Lowells time,
astronomical photography had come on the scene and offered a much
faster and more reliable way to survey the sky. Lowells rst search
was made between 1905 and 1907 using pairs of photographic plates
which he scanned by eye, placing one above the other and examining
them with a magnifying glass. He soon realised that this method was
not going to work.
Lowells next step was to order a device known as a blink compara-
tor to assist in the examination of the photographs. The comparator
provided a magnied view of a portion of the photographs but, more
importantly, it allowed the searcher to switch rapidly between two dif-
ferent images of the same patch of sky. Once the photographs were
T
h
e

e
d
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
5
aligned correctly, star images remained stationary as the view ashed
from one plate to the other. However, should there be a moving object
in the eld of view, its image would jump backwards and forwards as
the images were interchanged. Naturally enough, the process was
known as blinking the plates.
A search of the constellation Libra was made in 1911, but was aban-
doned after a year when nothing was found. Undeterred by this
failure, Lowell began another search in 1914. Between then and 1916
over 1000 photographic plates were taken, but once again nothing was
found.

Lowell died suddenly from a stroke on the 16th of November

1916, his planet-nding ambition unfullled. He was buried in a small
mausoleum, shaped to resemble the planet Saturn, in the grounds of
his observatory on Mars Hill. For a time the search for Planet X was
halted as Lowells widow tried to break the provisions of his will. Mrs
Lowell wanted to remove funds from the operation of the observatory
and preserve the site as a museum in her late husbands memory. The
resulting litigation siphoned off funds from the observatory for a
decade.
Eventually, under the directorship of Vesto Slipher, the Lowell
Observatory returned to the problem of the missing planet. Slipher
recruited a young amateur astronomer named Clyde Tombaugh, a
farmboy from Kansas, as an observing assistant. Tombaugh arrived
in Flagstaff during January 1929 and was set the task of taking pho-
tographs which could be searched for Lowells Planet X. It took a while
to get the new 31 cm telescope, built especially for the search, into full
operation, but by April all was ready. Tombaugh took a number of
photographic plates covering the region around the constellation of
Gemini, the latest predicted location of Planet X. The plates were
33.5 cm by 40 cm in size and covered nearly 150 square degrees of sky.
Each contained many thousands of star images. Vesto Slipher and his
brother blinked the plates over the next couple of weeks, but they
failed to nd anything. In the meantime, Tombaugh continued to pho-
tograph the sky and soon a large backlog of unexamined plates had
built up. Slipher then asked Tombaugh to blink the plates as well as
taking them, explaining that the more senior observatory staff were
too busy to devote much time to the onerous and time-consuming
blinking process.
B
e
y
o
n
d

P
l
u
t
o
6

In fact the missing planet did appear on two of these images, but it was much fainter
than expected and its presence was missed.
T
h
e

e
d
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
7
Figure 1.2 Clyde Tombaugh entering the dome of the Lowell Observatorys 33 cm tele-
scope. He is carrying a holder for one of the photographic plates. After exposing the plate
he had to search it millimetre by millimetre for Planet X. Few astronomers now go to their
telescopes so formally dressed, as can be seen by comparison with gures 4.1 and 5.7.
(Lowell Observatory archives.)
Tombaugh regarded the prospects of his new assignment as grim,
but he dutifully continued with his programme. Night after night he
made a systematic photographic survey of the sky. He concentrated
on regions close to the ecliptic, an imaginary line across the sky
which marks the path traced out by the Sun across the constellations
of the zodiac during the course of a year. The ecliptic is not the precise
plane of the solar system, which is better dened by taking account of
all the planets and not just of the Earth. When this is done the result is
known as the invariable plane. However, when projected onto the sky,
the ecliptic and the invariable plane are not much different and it is
common, if careless, to use the two terms interchangeably. Since the
orbit of Planet X would presumably be close to the invariable plane,
the ecliptic was the obvious region around which to search.
Tombaughs method was to take three photographs of each region
of sky at intervals of two or three days. Each photograph was exposed
for several hours. During each exposure Tombaugh painstakingly
guided the telescope to make sure that the images of the stars were
sharp, with all their light concentrated onto as small an area of the
photographic emulsion as possible. Only then would his plate reveal
the very faintest objects and give him the best chance of success. At
dawn he developed the plates, careful lest a tiny mistake ruin them
and waste his hours of work in the telescope dome. Later he examined
the plates for anything which might have moved between the two
exposures.
B
e
y
o
n
d

P
l
u
t
o
8
Figure 1.3 The orbits of Jupiter,
Neptune and Pluto. Plutos eccentric
orbit crosses that of Neptune,
although the signicance of this was
not realised at the time of its discov-
ery. (Chad Trujillo.)
Although the technique sounds simple in principle, Tombaughs
task was a huge one. The long nights of observing were tiring and the
blinking of the frames was tedious in the extreme. Many false detec-
tions appeared, caused by things such as variable stars, chance align-
ments of main belt asteroids and photographic defects which
mimicked moving objects. To eliminate these false detections,
Tombaugh used his third plate to check if any of the candidate objects
were visible again. Usually, of course, they were not. Tombaughs
patience was nally rewarded on the 18th of February 1930 when he
was examining a pair of plates he had taken a few weeks earlier.
Blinking them, he found a moving object that was clearly not a star, a
nearby asteroid or a aw in the photographic emulsion. What was
more, the objects slow motion across the sky suggested that it must be
well beyond Neptune. After a few more weeks of observations had
been made to dene the objects motion more accurately, the discovery
was announced on 13th March. The date was chosen since it would
have been Lowells 75th birthday if only he had lived to see the day.
After a certain amount of debate, to which we shall return later, the
new object was named Pluto, after the god of the underworld.
Clyde Tombaugh continued his search for another 13 years. He esti-
mates that in this time he covered about 70% of the heavens and
blinked plates covering some 90000 square degrees of sky.

All in all he
spent some 7000 hours scanning every square millimetre of about 75
square metres of plate surface. Although he marked 3969 asteroids,
1807 variable stars and discovered a comet, he never found another
object as distant as Pluto. This was a little odd since it gradually
became clear that the new planet was rather smaller than predicted.
The rst clue that Pluto was small came from its faintness, which sug-
gested it could not be any larger than the Earth. Worse still, even the
largest telescopes of the day could not resolve Pluto and show it as a
disc. Under even the highest magnications, the planet remained a
tiny point of light, devoid of any features. This was worrying since if
Pluto was very small it could not affect the orbit of Uranus to any
signicant extent. None the less, the intensity of Tombaughs efforts
seemed to rule out any chance that any other massive planet could
exist near Neptunes orbit.
It was not until much later that theoretical work, notably by
T
h
e

e
d
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
9

The total area of the sky is less than this, but some regions were examined more than
once.
American E. Myles Standish in 1993, explained the apparent devia-
tions in the motion of Uranus. Standish based his calculations on
improved estimates of the masses of the giant planets which had been
determined during the ybys of the Voyager spacecraft. Using these
he showed that any remaining errors in the measurements of Uranus
position were tiny and could be explained by systematic observational
uncertainties. There was no need to invoke the gravitational inuence
of a missing planet, massive or otherwise. Lowells hypothesis of a
Planet X had been completely wrong. The discovery of Pluto was a
consequence of the thoroughness of Tombaughs systematic search
and the fact that Pluto was fairly close to Lowells predicted position
was just a coincidence.
It was well into the 1970s before the true nature of Pluto was
revealed. The planets orbit was quite well dened within a year of its
discovery, but Plutos faintness made determining details of its physi-
cal make-up almost impossible for decades. In the mid 1950s it was
established that Pluto has a rotation period of 6.39 days and in 1976
methane frost was detected on its surface. Since methane frost is quite
reective, this implied that Pluto was even smaller than at rst
thought. Pluto soon shrank again. In 1977 James Christy was examin-
ing images of Pluto when he noticed that the planet seemed to be elon-
gated some of the time and not others. He soon realised that this was
due to the presence of a large satellite going around the planet every
6.39 days, the same as Plutos rotation period. As its discoverer,
Christy had to name the new moon and he chose Charon, the name of
the ferryman who transported souls to the underworld. Strictly speak-
ing Charon should be pronounced Kharon, but it is often enunciated
as Sharon since Christys wife, Sharlene, is known to her friends as
Shar. Once the details of Charons orbit had been established, it was
possible to determine the combined mass of Pluto and Charon. This
turned out to be no more than 0.0024 times the mass of the Earth.
Pluto was a small and icy world. Although the true size of Pluto was
unclear in the 1940s, it may have been the realisation that there was no
massive Planet X that made Edgeworth and Kuiper speculate about
the edge of the solar system. Certainly the existence of small icy
worlds at the fringe of the planetary region seemed a natural conclu-
sion from theories of how the solar system formed.
It had once been suggested that the solar system was produced
when a close encounter between our Sun and another star pulled out a
lament of material which condensed into planets. However, it was
B
e
y
o
n
d

P
l
u
t
o
10
soon shown that this could not be the case. The realisation that the dis-
tances between the stars were very large made such an encounter
unlikely, but more importantly, it can be shown mathematically that
material pulled out from the Sun could not form planets. Ejected mate-
rial would either fall back onto the Sun or disperse into space. So
astronomers rejected this near-encounter model. Instead, they
embraced an idea put forward by the philosopher Immanuel Kant in
1755 and subsequently developed by a French scientist, Pierre Simon,
Marquis de Laplace. In 1796 Laplace suggested that the Sun formed in
a slowly rotating cloud of gas and that, as the cloud contracted, it
threw off rings of material which formed the planets. Although many
of the details have been improved, the general outline of this nebular
hypothesis survives today.
Modern theories of the formation of our solar system begin from
the assumption that stars like the Sun form in the clouds of gas and
dust which exist throughout interstellar space. These clouds often
contain as much as a million times more mass than the Sun and each
spreads over a huge volume of space. From time to time, instabilities
develop within these clouds and bursts of star formation are trig-
gered. About ve billion years ago, an instability in just such a cloud
triggered one such collapse. At the centre of this collapsing region,
itself buried deep within the larger interstellar cloud, a dense clump
of material began to form. As this protostellar core contracted, it
increased in mass and so generated a more powerful gravitational
eld. This in turn attracted in more material, increasing the mass of
the core still further in a rapidly accelerating process. As material fell
in towards the centre it was slowed down by friction and gave up its
kinetic energy as heat, gently warming the central regions of the
core. For a while, the heat could leak out in the form of infrared and
sub-millimetre radiation and so the collapsing core remained quite
cool. However, as the cloud got more and more dense, a point was
reached when its central regions became opaque to most forms of
radiation. When this happened, heat could no longer escape easily and
the temperature at the centre began to rise rapidly. After a while con-
ditions reached the point at which nuclear reactions could begin and
the core began to convert hydrogen to helium. The energy released by
these nuclear reactions generated sufcient pressure to halt any
further collapse and the star we call the Sun was born.
Of course the details of the star formationprocess are far more
complicated thancanbe described ina single paragraph. Inparticular,
T
h
e

e
d
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
11
a mathematical analysis of the fate of a spherical collapsing cloud
immediately throws out a simple, but vitally important question. If
the Sunformed fromthe collapse of a huge cloud of gas, why does it
rotate so slowly, taking about 11 days to turnonits axis? This fact alone
hints at the existence of planets as a consequence of the physical law
that angular momentum, or spinenergy, must be conserved.
The conservation of angular momentum can be observed when an
ice dancer skating with arms outstretched enters a tight turn and
begins to spin on the spot. If, as her spin begins to slow down, the
dancer brings her arms in close to her body, her rate of rotation sud-
denly speeds up. A similar effect can be experienced, without getting
cold feet, by sitting on a well oiled ofce chair and spinning around on
it with your arms held out. If you pull in your arms you can feel the
spin rate increase, push them out again and the spinning slows down.
Try again with a heavy book in each a hand and you will nd it works
even better. This simple observation is revealing two important things
about physics. Firstly, angular momentum depends on both the rate at
which something is spinning and upon the distance of its constituent
masses from the axis of rotation. Secondly, the total amount of
angular momentum in a spinning system is conserved. So, as demon-
strated by our ice dancer, as mass is brought in towards the axis of
rotation of a spinning system, the spin rate must increase to keep the
total amount of angular momentum, or spin energy, the same. The
more mass there is on the outside of the spinning region, and the
further the mass is from the spin axis, the more angular momentum
the system has.
The problem faced by the forming Sun was that as the protostellar
cloud collapsed, it had to lose considerable amounts of angular
momentum. This is necessary because unless the original cloud was
completely at rest when the infall began, then as material fell inwards,
it would have transferred its angular momentum to the central
regions. This would have increased the rate of rotation of the proto-
sun quite dramatically. Unless this angular momentum could be
removed, the spin rate would continue to increase as the collapse pro-
ceeded. By the time the core had shrunk to stellar dimensions, the
rotation would be far too rapid to allow a star to form. So, somehow
during its collapse, the core must have transferred angular momen-
tum to material further out in the cloud. This occurred as magnetic
elds and gas drag gradually forced the outer reaches of the cloud to
spin around with the core. As this continued, the outer regions of
B
e
y
o
n
d

P
l
u
t
o
12
what had been a spherical cloud fell in towards the equator and the
nebula became a huge, slowly spinning disc surrounding a small
stellar embryo. The forming Sun continued to grow as material in the
disc fell inwards onto it.
The conditions across the protoplanetary disc depended on the
balance between the energy generated during the collapse, the light
emitted by the still-forming Sun and the rate at which energy was
transported through the disc. In the central regions it was too hot for
icy material to survive. Here, in what became the inner solar system,
most of the ices were evaporated and blown outwards, leaving behind
more robust dusty material. Within the spinning disc, tiny grains
began to bump into each other. The grains, remnants of the original
interstellar cloud, were probably smaller than a micron across to start
with, but the collisions were gentle enough that they began to stick
together. At rst they formed uffy structures which were mostly
empty space, but as they grew still further, they began to compact.
Soon they reached the point were they were more like small pebbles,
jostling each other as they orbited the Sun. Inexorably these lumps of
debris grew still further. Then, once a few objects had reached a size of
about ten kilometres in diameter, a dramatic change of pace occurred.
These larger lumps, or planetesimals, were now massive enough
that their gravitational elds began to attract other passing material
onto themselves. Once this started it dramatically accelerated the
growth process. Before long a few planetesimals began to dominate all
of the space around them, clearing away the remainder of the orbit-
ing material by dragging it down onto their surfaces. Within 100000
years or so many rocky bodies about the size of the Earths moon had
formed. After this brief period of runaway growth, the pace slowed
again. By now each planetary embryo had swept up all the material
within reach and the distances between the larger objects were too
great for them to encounter each other. It took another 100 million
years for the planet-building process to be completed. Gradually,
subtle gravitational interactions between the planetesimals stirred up
their orbits enough for occasional dramatic collisions to occur. One by
one the surviving embryos were swept up into the four terrestrial, or
Earth-like, planets, which we see today.
Further out, about half a billion kilometres from the Sun, tempera-
tures were low enough that ices could survive. So, as well as dust, the
outer regions of the disc contained considerable amounts of water ice
and frozen gases such as methane, ammonia and carbon monoxide.
T
h
e

e
d
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
13
Here, the growing planetary cores swept up this extra material to
form giant planets dominated by the gases hydrogen and helium with
a seasoning of various ices. Jupiter, the largest of these giants, was so
large that, even while it was still forming, its gravitational eld had a
dominating effect on its neighbourhood. Jupiters gravity stirred up
the region between itself and the still forming planet Mars and pre-
vented a single object dominating this region. Instead of forming a
fully edged planet, the growth stopped, leaving a population of
smaller, rocky asteroids. Beyond Jupiter, the other giant planets
Saturn, Uranus and Neptune grew as they too swept up the icy plan-
etesimals from the space around them.
At great distances from the Sun the protoplanetary disc became
much more diffuse. Here, although there was sufcient material to
reach the stage of forming small planetesimals, there was not enough
time, or enough material, for them to combine into a major planet.
Instead they formed a diffuse zone of small icy objects in almost per-
manent exile at the fringes of the solar system. This is the frozen
boundary of the planetary region; beyond it lies only the huge, more-
or-less spherical cloud of planetesimals ejected into deep space by
gravitational interactions with the forming planets and the rest of the
stars in our galaxy.
After Edgeworths and Kuipers articles, thinking about a possible
disc of planetesimals beyond Neptune lapsed until the early 1960s.
Abrief revival of interest began in 1962 when naturalised American
physicist Alistair Cameron

wrote a major review about the formation

of the solar system. Camerons review appeared in the rst issue of
Icarus, a new scientic journal devoted exclusively to the study of the
solar system. Using the same arguments as Kuiper and Edgeworth,
namely that material in the outer regions of the protoplanetary disc
would be too diffuse to form a planet, Cameron wrote that It is
difcult to escape the conclusion that there must be a tremendous
mass of small material on the outskirts of the solar system. Soon
after Cameron wrote his review, another astronomer turned his atten-
tion to the possible existence of a comet belt beyond Neptune.
American Fred Whipple, who had done much to explain the compo-
sition of comets a decade earlier, began by accepting the likely exis-
tence of what he called a comet belt. From his knowledge of comets,
he reasoned that if a trans-Neptunian belt of icy planetesimals
B
e
y
o
n
d

P
l
u
t
o
14

Cameron was originally a Canadian.

existed, then even objects as large as 100 km in diameter would be very
faint. This would make the discovery of individual objects highly
unlikely with then existing astronomical technology. Turning the
question around, he then asked himself if the comet belt would be
detectable if the comets within it were very small. Would the com-
bined light of large numbers of small comets produce a faint, but
detectable glow across the sky? His conclusion was that any glow from
the comet belt would be too faint to see against the background of the
night sky. In particular it would be masked by the diffuse glow of the
zodiacal light, a band of light along the ecliptic plane produced by sun-
light scattering off interplanetary dust in the inner solar system.
Having decided that it was impossible to detect a hypothetical comet
belt directly, he set out to attack the problem dynamically. Harking
back to Adams, Le Verrier and Lowell, Whipple tried to nd out if a
comet belt could have any measurable gravitational effects on the rest
of the solar system.
Whipple rst considered the gravitational effect of the belt on the
motion of the planets Uranus and Neptune. He concluded that a comet
belt having a mass of 1020 times that of the Earth might exist beyond
Neptune, but that the evidence for such a belt was not conclusive. He
merely noted that a hypothetical comet belt provided a better explana-
tion of the apparent irregularities in the motion of Neptune than
assigning a mass to Pluto that was much larger than seemed justied
by other observations of the tiny planet. He even went as far as to say
Pluto could not affect the other two planets signicantly even if it
were made of solid gold. In 1967 Whipple, together with S.E. Hamid
and a young astronomer called Brian Marsden, tried to estimate the
mass of the comet belt another way. They looked for its effect on the
orbits of seven comets which all travelled beyond Uranus. They then
calculated the gravitational effect that a hypothetical comet belt con-
taining as much material as the Earth would have on the orbits of
each of these comets. Since they found that the real comets had suf-
fered no such effects, they concluded that any unseen comet belt
could not have a mass of much more than one Earth mass. Thus
Edgeworths and Kuipers ideas remained largely in limbo for a
number of years. It was only when a number of advances in our
understanding of comets began to come together that it was gradually
realised that there was a problem that could best be solved by postulat-
ing the existence of an ecliptic comet belt.
T
h
e

e
d
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
15
The existence of comets, as ghostly apparitions that appear
without warning, move slowly across the sky and then fade away, has
been known throughout history. However, only in the latter half of the
twentieth century was a reasonable physical model of a comet devel-
oped. Although Edmund Halley noticed the similarity between the
orbits of a number of comets, realised they were the same object and
predicted the return of what has become known as Halleys Comet,
neither he nor his contemporaries really understood what a comet
actually was. By the early 1900s the favoured model was of a loose
aggregation of dust and rocks, little more than a loosely bound cloud
of material, carrying with it gas molecules trapped both on the sur-
faces of the grains and deep within pores of the larger pieces. When
the comet was warmed by the Sun, these gases were apparently
released to form a tail. There were serious aws with this model, the
most signicant being that such a system could not supply enough gas
to explain the rate at which gas was known to be produced as a comet
approached the Sun. There was really only one thing that was known
for certain about comets: dynamically speaking, they were of two dis-
tinct types. Comets of one kind appeared unpredictably from random
directions on the sky and made a single trip around the Sun before dis-
appearing for thousands of years. Those of the other kind, which were
generally much fainter, reappeared regularly and their returns could
be predicted quite accurately.
Comets of the rst kind, called long-period comets, follow very
elongated (parabolic) orbits which range from the inner solar system
at one extreme into deep space at the other. As with most solar system
objects, it is convenient to describe these orbits in terms of astronomi-
cal units (AU). An astronomical unit is dened as the average distance
of the Earth from the Sun and amounts to about 150 000 000 km. Using
these units, Jupiter is 5 AU from the Sun and Neptunes orbit is at
about 39 AU. The long-period comets which can be observed from
Earth have perihelia, or closest approaches to the Sun, of less than a
few astronomical units and aphelia, or furthest distances from the
Sun, of many thousands of AU. In the late 1940s the Dutch astronomer
Jan Oort examined the statistics of the few hundred long-period
comets then known and suggested that they came from a huge, more-
or-less spherical shell around the Sun which extended to about half
way to the nearest stars. Oort believed that the comets were ancient
planetesimals that had been gravitationally ejected from the region of
what is now the asteroid belt during the planet-building process about
B
e
y
o
n
d

P
l
u
t
o
16
four and a half billion years ago. He suggested that they then remain
in the distant cloud until random gravitational forces from other
nearby stars change their orbits slightly and cause them to fall
inwards towards the warm heart of the solar system. Gerard Kuiper,
an ex-student of Oorts, soon pointed out that the comets, being icy,
were probably formed in the region from 35 to 50 AU rather than in the
asteroid belt. Kuiper believed they were ejected by Pluto, not Jupiter.
However, the broad outline of Oorts theory for the origin of comets
was generally accepted and the hypothetical shell of distant comets
became known as the Oort Cloud.
About the same time as Oort was explaining the dynamics of the
long-period comets, Fred Whipple brought forward his icy conglomer-
ate or dirty snowball model of a comet. He suggested that the essence
of a comet was a single solid body a few kilometres across called the
nucleus. Each comet nucleus comprises frozen ices such as water,
carbon monoxide, ammonia and methane together with a small
amount of dust. As the nucleus approaches the Sun, solar heating
warms it and causes the frozen gases in its outer layers to sublime.
This creates a physically large, but very tenuous cloud around the
nucleus. This cloud is called the coma. The coma is not entirely gas,
since as the gases leave the nucleus they carry with them tiny dust
particles. The pressure of sunlight, and the solar wind of material
T
h
e

e
d
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
17
Figure 1.4 Comet orbits. Long-
period comets approach the Sun
along parabolas. Short-period comets
orbit the Sun in ellipses usually, but
not always, quite close to the plane of
the planets. A comet on a hyperbolic
orbit would be approaching from
outside the solar system. The fact
that no such hyperbolic comets are
seen is evidence that comets are part
of the Suns family. The elliptical
orbit has an eccentricity of 0.9.
constantly owing out from the Sun, act on the coma and blow mater-
ial away to form the comets tail. Most comets actually have two tails, a
long straight bluish one comprising gases that have been ionised and
are moving directly away from the Sun and a curved, yellowish one
comprising individual dust grains being blown away from the nucleus
into independent solar orbits. In most comets, depending on the ratio
of gas to dust in the nucleus, one of these tails is much more promi-
nent than the other. Sunlight reected from the coma and the dust tail
makes the comet visible from the Earth. Once the comet has passed
around the Sun and begins to recede back into deep space, the nucleus
cools and the sublimation of the ices slows down and nally stops.
Once this happens the comet rapidly becomes too faint to detect.
Unseen, the frozen nucleus returns to the Oort Cloud from where,
thousands of years hence, it may return to visit the Sun again.
The second class of comets are those of short period. These are
conned to the inner solar system and most of them travel around the
B
e
y
o
n
d

P
l
u
t
o
Figure 1.5 Comet HaleBopp seen from Mauna Kea, Hawaii. Comet HaleBopp is a long-
period comet and displayed a long, bright dust tail. Short-period comets are almost never
visible to the unaided eye. (John Davies.)
Sun in elliptical orbits with periods of about a dozen years. In general,
they have orbits of low inclination, that is to say they are close to the
plane of the solar system. The short-period comets are generally
much fainter than comets coming from the Oort Cloud. This is
because short-period comets approach the Sun very frequently and on
each trip more and more of the volatile ices which form the coma and
tail are removed. So even when heated by the Sun at perihelion, short-
period comets are pale shadows of their fresh, bright cousins making
their rare appearances in the inner solar system. The faintness of the
short-period comets indicates that they are gradually running out of
volatile material and that they cannot survive for long in their present
orbits. The short-period comets are fated to fade away completely, and
to do so quite quickly in astronomical terms. By estimating how much
material is removed on every trip around the Sun, astronomers have
shown that short-period comets cannot survive in their present loca-
tions for even a small fraction of the age of the solar system. However,
the very fact that numerous short-period comets do exist means that
new ones must be arriving regularly to top up the present supply and
replace them as they vanish. Many of these short-period comets have
aphelia in the region of Jupiters orbit; these are called Jupiter family
comets. This link with the giant planet is a clue to their origin. Comets
from the Oort Cloud which happen to approach Jupiter too closely
have a chance of being captured into the inner solar system by
Jupiters gravity, making them doomed to make frequent approaches
to the Sun until they disappear forever.
Although the Dutch astronomer Van Woerkom had noticed in 1948
that there seemed to be about twenty times more short-period comets
than he would have expected, the idea that short-period comets were
really just ordinary comets from the Oort Cloud which had the misfor-
tune to be captured by Jupiter was accepted for a number of years.
However, as more and more comets were discovered, it became clear
that something was wrong. The observed population of short-period
comets was too large to be explained by the effects of Jupiters gravity
on comets from the Oort Cloud. The developing problem was twofold.
Firstly, the capture of an individual comet by Jupiter is a very
unlikely event. Only if a comet ies quite close to Jupiter can enough
gravitational energy be exchanged to slow the comet down and trap it
in the inner solar system. With comets arriving from all directions,
including from above and below the plane of the planets, the chances
of crossing Jupiters orbit just when the planet happens to be there,
T
h
e

e
d
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
19
and to do so close enough to the plane of the solar system to be cap-
tured, are very small. The probability of capture in this way is so low
it has been compared with the likelihood of hitting a bird with a
single bullet red into the sky at random. Also, most of the short-
period comets are in low-inclination orbits and most go around the
Sun in the same sense as the rest of the planets. Since comets from the
Oort Cloud approach the Sun at all angles to the ecliptic plane, includ-
ing in orbits that go around the Sun in the opposite direction to the
planets, it was puzzling that orbits of the short-period comets were
not more randomly distributed.
In 1972, physicistastronomer Edgar Everhart tried to resolve this
problem. He suggested that the short-period comets were derived not
from the capture of just any Oort Cloud comet, but rather from a
subset of such comets which had specic orbital characteristics that
made them likely to be captured. In particular, Everhart suggested
that the short-period comets were Oort Cloud comets which had
entered the zone of 46 AU from the Sun close to the plane of the solar
system. Everharts model could explain why the short-period comets
population looked the way it did, but it was not long before another
problem showed up. The following year Paul Joss from Princeton
University looked at Everharts model and put in some estimates for
the capture rate, the likely lifetime of a typical short-period comet,
and so on. In just two pages of text he showed that there were hun-
dreds of times too many short-period comets to be explained by
Everharts capture model. Of course, not everyone agreed with him,
but it did look as if something was missing from the equation. While
the dynamicists pondered this problem, a new piece of the puzzle was
about to be turned up by a strictly observational astronomer.
B
e
y
o
n
d

P
l
u
t
o
20
The Centaurs
The 1970s was a golden age for solar system exploration. Robot explor-
ers orbited Mars, landed on Venus and ew sunwards to photograph
the innermost planet, Mercury. In 1976, two Viking spacecraft landed
on Mars to search for signs of life and, as the decade drew to a close,
Voyagers 1 and 2 visited the gas giant planets Jupiter and Saturn. As
mission followed mission, thousands of stunning images ooded back
to Earth. One by one a dozen or more planets and moons were trans-
formed from blurry images or points of light into individual worlds
with distinct personalities. Across the solar system, newly discovered
mountains, valleys and craters were mapped, catalogued and then
named. Almost overnight planetary science moved from the realm of
astronomy to become more akin to geology and geography. However,
despite the huge amounts of data being returned from space, a few
areas of solar system research remained the province of traditional
ground-based telescopic observers. With space missions concentrat-
ing on exploring the planets and their satellites, the rest of the solar
system seemed to be something of a backwater. The numerous comets
and asteroids lacked glamour and received comparatively little atten-
tion. None the less, a tiny band of astronomers struggling to make
sense of the population of small bodies was beginning to make some
progress in understanding the structure and composition of the main
asteroid belt. As the Voyager missions revealed that much of the outer
solar system was dominated by icy material and that large impact
craters were found throughout the solar system, a few people began to
ask questions about what else might lurk in the huge volume of space
beyond Jupiter.
Although detailed thinking about the trans-Neptunian region had
hardly begun at this time, the idea that a population of hitherto
C
H
A
P
T
E
R

2
21
unknown planetesimals might exist received a boost in 1977 with the
discovery of an unusual new member of the solar system. The object
was found by Charles Kowal, a native of Buffalo, New York State.
Always interested in an astronomical career, Kowal moved to
California when he was sixteen and took a degree in astronomy at the
University of Southern California. Tired of working 40 hours a week
to support himself while studying, he decided not to stay on to gain a
higher degree and, as the end of his studies approached, he wrote to
various observatories in search of a job. After considering several
possibilities he eventually accepted a position as a research assistant
at the Mount Wilson and Mount Palomar Observatories, then home of
some of the worlds biggest telescopes. During this period Kowal
worked for such well-known astronomers as Allan Sandage and Fritz
Zwicky. He became interested in the solar system because asteroid
trails frequently appeared by chance on the photographs of distant
galaxies which he was taking for Zwicky. Around 1970, Kowal started
searching for asteroids whose orbits brought them close to the Earth
and he soon realised that the 1.15 m Schmidt telescope on Mt Palomar,
which he had used during observations for Fritz Zwickys projects,
would be an ideal tool to search for new satellites of the outer planets.
The Schmidt telescope, essentially a huge camera, was capable of
taking photographs covering an area of sky 6 degrees 6 degrees and
a single plate would encompass the whole satellite system of a planet
like Jupiter. Using this telescope, Kowal discovered Jupiters thir-
teenth satellite in 1974. He named the new satellite Leda. In 1976, he
expanded his work to begin a systematic photographic survey aimed
at detecting hitherto unknown distant solar system objects.
For his search, Kowal used a photographic blinking technique
similar to the one used by Clyde Tombaugh four decades earlier. Like
Tombaugh, Kowal took a series of exposures concentrated on posi-
tions spread along the ecliptic and then blinked them to search for
moving objects. On 1st November 1977 he was examining a pair of
plates which he had exposed two weeks earlier when he found a faint
object whose slow motion across the sky, equal to about 3 arcminutes
per day, suggested that it might be about as distant as the planet
Uranus. Following the established convention, Kowal reported his
nd to the International Astronomical Unions (IAU) Central Bureau
for Astronomical Telegrams.
The Central Bureau for Astronomical Telegrams, known some-
times as simply CBAT and occasionally by the rather more Orwellian
B
e
y
o
n
d

P
l
u
t
o
22
name of the Central Bureau is housed at the Smithsonian
Astrophysical Observatory in Cambridge, Massachusetts. Its director
was Brian Marsden, an avuncular, expatriate Englishman who had
developed an interest in cometary orbits as a schoolboy in the 1950s.
After getting a degree from Oxford University, Marsden left England
for Yale University and enrolled in that establishments respected
celestial mechanics programme. He was awarded a PhD in 1965 and, at
the invitation of Fred Whipple, moved to the Smithsonian
Astrophysical Observatory. Three years later he became director of
the Central Bureau, responsible for the issuing of IAU telegrams and
circulars. In the days before the internet, IAU telegrams were used to
alert astronomers to potentially interesting discoveries such as
comets, variable stars, supernovae and other rapidly changing phe-
nomena. Today, the telegram service has been retired in favour of
electronic mail, although printed copies of announcements still go
out by regular mail as postcard-sized IAU circulars.
Marsden shared the news of Kowals nd with the Minor Planet
Center, a sort of central clearing house for observations of comets and
asteroids. Originally, centres for the calculation of asteroid and comet
T
h
e

C
e
n
t
a
u
r
s
23
Figure 2.1 Charles Kowal at the
blink comparator which he used to
discover the rst Centaur, Chiron.
(California Institute of Technology.)
orbits had been concentrated in Europe, with much of the work being
done in Germany. Predictions of the positions of known asteroids
appeared in the annual publication Kleine Planeten and newly discov-
ered objects were announced via the circulars of the Rechen-Institut.
These arrangements collapsed completely during the Second World
War. After the war the Kleine Planeten was taken over by the Institute
of Theoretical Astronomy in Leningrad and became the annual publi-
cation Efemeridy Malyth Planet. The work of the Rechen-Institut was
transferred to a new Minor Planet Center established in 1947 by Paul
Herget in Cincinnati, Ohio. The purpose of the Minor Planet Center
was to receive observations and publish circulars containing accurate
positions of newly discovered comets and asteroids. The center
remained in Cincinnati until Hergets retirement in 1978, at which
time it moved to Cambridge, Massachusetts, and came under
Marsdens wing. Until he retired in August 2000, Marsden headed a
small staff of experts in the esoteric eld of orbital mechanics. The
group, now headed by Dan Green, record observations of solar system
objects and circulate the details to other interested parties, so that
new sightings can be conrmed and followed up in a timely fashion.
Once sufcient observations have been made, the staff of the
B
e
y
o
n
d

P
l
u
t
o
24
Figure 2.2 Charles Kowals discovery image of Centaur 2060 Chiron. The trail is much
shorter than the fainter one of a main belt asteroid at the upper right edge. This indicates
that Chiron is moving more slowly, and so is more distant, than a typical asteroid. (Charles
Kowal.)
Minor Planet Center calculate and then publish the orbits of the
newly discovered bodies. At this point the Director of the Minor
Planet Center has the sometimes delicate task of attributing credit for
the discovery. Comets are usually named after their discoverer, or dis-
coverers, although there are a few exceptions to this rule. Comets
Halley and Crommelin are named after the individuals who calcu-
lated the details of their orbits rather than the astronomers who rst
observed them, and comets which are discovered using satellites or
automated search telescopes are named after the project, rather than
the people involved. Thus there are numerous comets called Solwind,
IRAS, SOHO and LINEAR. Unlike comets, asteroids initially receive a
temporary designation until their orbits have been calculated with
sufcient precision that they can be found at any time in the foresee-
able future. Once this has been done, the object is assigned a perma-
nent minor planet number and the discoverer is invited to suggest a
name for the new object.
Since photographs of Kowals new object showed no sign of a
comet-like coma, it was given the provisional asteroid designation of
1977 UB. The designation followed a simple code in use since 1925
which provides information on the approximate time of discovery.
This designation is used until a permanent name and number can be
assigned.

The rst part of the designation is simply the year of dis-

covery and the second part denes when in the year the object was
rst seen. The rst half of the rst month of the year is designated A,
the second half B and so on throughout the alphabet, ignoring the
letters I and Z. The second letter is assigned in the order that reports
are processed by the Minor Planet Center. So the rst object reported
in the rst half of January 1977 was known as 1977 AA, the second
1977 AB and so on until the middle of the month when the designa-
tions 1977 BA, 1977 BB, 1977 BC started to be applied. Kowals object
was the second one reported in the second half of October and so it
was designated 1977 UB.
Luckily, another astronomer, Tom Gehrels, had observed the same
area of sky on the 11th and 12th of October. An examination of his
photographs soon revealed the new asteroid near the corner of
his photographs. With these extra positions, together with other
T
h
e

C
e
n
t
a
u
r
s
25

Until 1925 there had been three previous cycles (18931907, 19071916 and 19161924)
during which pairs of letters were assigned sequentially without regard to the time of
year.
observations made a few days later, it was possible for Brian Marsden
to calculate rough details of the new objects orbit. These showed that
1977 UB was indeed distant from the Sun, but that it did not appear to
be a comet entering the solar system for the rst time. The initial cal-
culations suggested that it was in a roughly circular orbit between the
planets Saturn and Uranus. If this was true, then it represented a
completely new class of solar system object and there was much inter-
est in trying to determine the orbit more accurately. To do this,
astronomers used early estimates of the orbit to project the objects
motion back in time. They then hunted through their libraries of
photographs to see if there was any chance that it might have been
recorded accidentally on images taken for some other purpose. Since
the new object was fairly bright, and so quite easy to nd once one
knew roughly where to look, Kowal was able to nd images on two
photographs he had taken in September 1969. With this new informa-
tion, more images were soon found, some from 1976 and some from
August 1952. A new orbit derived from these observations allowed
William Liller to locate the object on a plate taken in 1941 with a 61 cm
telescope at the Boyden Observatory in South Africa. A number of
other images were also found, including ones from 1943, 1945 and 1948.
There was even one on a plate taken in the USA in 1895 during pre-
delivery testing of the Boyden Observatory telescope.
When an object is located based on a projection of its orbit into the
future, the observation is called a recovery so, following the rather
diabolical practice of lm makers and authors in producing prequels
to popular lms and novels, observations made by projecting an orbit
backwards in time are called precoveries. Searches for precoveries are
quite common when interesting moving objects are discovered. This
is because old observations can be very useful in rening the objects
orbit, particularly when the object is very distant from the Sun and so
moving only slowly across the sky. When observations span only a
small period of time, the objects apparent motion, known as its
observed arc, will be small. When observations covering this arc are
used to calculate the objects orbital parameters quite large errors can
result. By providing a much longer baseline of observations, or a
longer arc, old observations make possible a more precise determina-
tion of the orbit, which in turn allows better calculations to be made
of the objects past and future positions. These improved estimates of
its past position sometimes make further precoveries possible, and
this improves the knowledge of the orbit still further. In the case of
B
e
y
o
n
d

P
l
u
t
o
26
1977 UB the precoveries soon covered a timebase of over 80 years. This
provided a big enough arc that its orbit was soon dened well enough
that it could be numbered minor planet 2060. At this point, the Minor
Planet Center, which acts under the authority of the International
Astronomical Union, asked Kowal to give the new asteroid a name.
The tradition of asteroid naming is a long one. It began in the early
1800s, when the Italian astronomer Piazzi named the rst asteroid
Ceres. The next two hundred or so asteroids also received classical
names, but with the introduction of photographic techniques for
asteroid hunting in the 1890s, and a consequent surge in new discover-
ies, the naming protocol began to be relaxed. Soon a variety of people,
places and organisations started to nd a place in the heavens. In the
case of obscure main belt asteroids the names chosen can sometimes
be quite frivolous. Minor Planet 2309 Mr Spock, which is named after
a cat, who was in turn named after the TV character, is a legendary
example. Fortunately, Kowal, who describes himself as One of those
old fashioned people who think that asteroids should be named care-
fully and who perhaps suspected that there was more to 1977 UB than
met the eye, was more circumspect. Since 1977 UB had been found
early in his survey, and hopeful of nding some more objects later,
Kowal looked for a group of mythological characters unrepresented
amongst the asteroids. He found that the Centaurs, strange creatures
that were half human and half horse tted the bill. From dozens of
Centaurs named in ancient literature, Kowal chose the name Chiron
(pronounced Kai-ron), the most prominent Centaur and arguably the
one with the best reputation. Mythologically speaking the Centaurs
were a rowdy bunch given to drinking, rape and pillage, but Chiron
was said to have devoted his efforts to astrology, medicine and the arts.
He is described by some scholars as the King of the Centaurs and was
known as a teacher as well as a healer. The choice was highly appro-
priate because the object was in an orbit between Saturn and Uranus
and mythologically speaking, Chiron was the son of Saturn and
grandson of Uranus. Although Kowal could not have known it at the
time, it was an inspired choice.
The discovery of 2060 Chiron raised several interesting questions
which occupied the popular media until the initial excitement died
down. One question that was asked, and that we shall encounter again
later in a different context, was whether Chiron was a new planet.
Although the initial observations did not reveal anything about
Chirons surface, making it impossible to be sure if it was covered in
T
h
e

C
e
n
t
a
u
r
s
27
reective material like ice or darker material such as rock and dust,
it was clear from its faintness that Chiron could not be more than a
couple of hundred kilometres across. Since there are a number of
main belt asteroids bigger than this, it was clear that Chiron was not
a planet in the traditional sense of the word. However, classifying
Chiron as a minor planet begged the question of whether it was
unique. Was it just the brightest member of a new trans-Saturnian
asteroid belt? Soon, further calculations based on the larger arcs
obtained by using the precovery observations were available. These
showed that the orbit of Chiron was not the circle originally esti-
mated, but was an ellipse ranging from inside Saturns orbit to a point
just inside the orbit of Uranus. Such an orbit is not stable and it was
clear that Chiron could only remain a denizen of the SaturnUranus
region for, at most, a few million years. Although neither its precise
past nor its eventual future could be calculated, it was easy to show
that Chiron will eventually approach close to either Saturn or Uranus.
When this happens gravity will drastically change its orbit, either
moving it further into space or perhaps sending it closer to the Sun.
Whatever its ultimate fate, Chiron is only a temporary resident of the
outer solar system.
Around the time of its discovery, it was variously suggested that
Chiron could have been an escaped satellite of one of the outer
planets, a rocky asteroid somehow ejected from the main asteroid belt
or perhaps a giant comet. The comet theory drew a parallel with
another unusual asteroid, 944 Hidalgo, which although rather smaller
than Chiron was also in an unusually eccentric orbit. Hidalgo was
thought by some astronomers to be a comet from which all the water
ice had been removed and which as a consequence was no longer
active. However, since in its present orbit Chiron does not approach
the Sun closely enough to sublime any water ice on its surface,
Marsden and co-workers suggested that Chiron was not so much a
dead comet, but rather one which had never lived. If they were right
and Chiron was a comet, it was a big one; its brightness suggested that
it was about fteen times bigger than the nucleus of Halleys comet.
A few more details of the nature of Chiron were revealed a decade
later when David Tholen of the University of Hawaii, amongst others,
showed that Chiron was brightening faster than expected as it
approached the Sun. Like planets, asteroids do not shine by them-
selves, they merely reect sunlight and their brightness at any given
time depends on a number of factors. The main ones are the size of
B
e
y
o
n
d

P
l
u
t
o
28
the object, how reective it is and its distance from both the Sun and
the Earth. Although it was not clear how reective Chiron was, and
therefore what its absolute brightness should be, it was fairly easy to
calculate how its brightness should change as it moved around the
Sun. Tholen and co-workers Dale Cruikshank and William Hartmann
found that Chiron was not sticking to the rules; throughout the late
1980s it was getting too bright to be explained by just its steadily
decreasing distance from the Sun. An airless objects reectivity, or
albedo, is determined by its surface composition and this is unlikely
to change dramatically over a period of a few years, so the most
obvious explanation was that the extra brightening was due to Chiron
developing a comet-like coma of gas and dust. A coma would drasti-
cally increase the area of material reecting sunlight and could cause
the anomalous brightening. These speculations were conrmed in
1989 when Karen Meech and others took images which showed that
Chiron had indeed developed a coma, and even had a comet-like tail
of material blowing away from the Sun.
At rst, this cometary outburst was ascribed to Chiron warming
up as it approached the Sun. It was speculated that solar heating
somehow caused gases trapped below the surface to blow a hole in an
insulating crust and allow a cloud of gas and dust to escape. Although
water ice would be expected to remain frozen at the distance of
Chiron, other volatile gases such as carbon monoxide, nitrogen and
methane could be responsible. These gases can be trapped in amor-
phous (water) ice and can be released if the ice undergoes a change to
the more regular, and familiar, crystalline form of ice. This transition
from amorphous to crystalline ice can occur at quite low tempera-
tures and has been proposed to explain the activity of various comets
at great distances from the Sun. However, a careful study of old pho-
tographs showed that Chirons activity was not restricted to periods
when it was relatively close to the Sun. Outbursts were detected even
when Chiron was near the most distant point of its orbit. To confuse
things still further, Chirons activity did not continue as it approached
the Sun and actually diminished or even stopped during its perihelion
passage in the 1990s. Whatever the source of the outbursts, it seemed
that Chiron looked like an asteroid some of the time, but like a comet
the rest of the time. Just like the mythical Centaurs, it was neither
one thing nor the other.
Charles Kowal continued his search for about eight years, nally
nishing in February 1985. By then he had observed 160 elds
T
h
e

C
e
n
t
a
u
r
s
29
totalling 6400 square degrees of sky. Although he found four comets
and several Earth-approaching asteroids, Kowal never did nd
another Centaur. Chiron remained a lonely enigma for almost fteen
years. The next step forward came from a project initiated by Tom
Gehrels, who had himself unknowingly recorded Chiron in 1977. The
Spacewatch project was established by Gehrels to make a long-term,
systematic search for new solar system objects using electronic detec-
tors instead of photographic plates, and computers instead of blink
comparators. From a telescope on Kitt Peak in Arizona, Spacewatch
was scanning the skies for several nights a month, making repeated
observations of the same area to nd objects that moved noticeably in
a few hours. Most of Spacewatchs discoveries had been of asteroids
B
e
y
o
n
d

P
l
u
t
o
30
Figure 2.3 An image of Chiron taken with a CCD camera in the 1980s. The telescope has
been tracked to allow for Chirons motion, making the stars appear as streaks. Despite
this, Chiron does not appear pointlike; the faint haze around it is evidence for a cometary
coma. Images like this conrmed that Chiron is a giant cometary nucleus. (Karen Meech.)
close to the Earth but, on 9th January 1992, David Rabinowitz was in
the Spacewatch control room when the systems moving-object detec-
tion software drew his attention to an object apparently moving too
slowly to be a normal asteroid.
Rabinowitz knew that false alarms resembling faint slow-moving
objects were quite common. From time to time, the software linked
together marginal detections of stars and electronic noise and
mistook them for a single object moving across the images. However,
Rabinowitz immediately realised that this source was much brighter
than a typical false alarm. A quick examination of the images showed
that the new object was pointlike, conrming its likely reality and
suggesting that it could be a new, distant asteroid. Rabinowitz phoned
Beatrice Mueller, who was working at the nearby 2 m telescope and
she immediately agreed to try and observe the new object. She made
the observations the same night. Within a few days, further observa-
tions had been made by the Spacewatch telescope and additional sets
of positions were being reported by other astronomers.
One observation was from a pair of plates which had been taken on
New Years Eve by the comet-hunting team of Gene and Carolyn
Shoemaker and their colleague David Levy. This group was really
searching for fast-moving objects, but on this night their photographs
also contained something that was moving rather slowly. They only
had observations on a single night, not enough for the Minor Planet
Center to do much more than le the positions away for future refer-
ence and certainly not sufcient for the object to be recorded as a pos-
sible new discovery. Because of this Carolyn Shoemaker admits she
did not give measuring the objects position very high priority. At the
time she thought it was probably a more-or-less ordinary asteroid,
unlikely to be followed up by anyone else and so probably fated to be
lost again. The observations were eventually reported to the Minor
Planet Center on the 13th of January. They arrived about the same
time as reports of another independent discovery made from photo-
graphic plates taken on 9th and 10th of January by Eleanor Glo
Helin. Well known for her work on asteroids passing close by the
Earth, Helin was observing from Mt Palomar as part of a long-
running search for fast-moving objects. As usual, while at the tele-
scope she was concentrating on scanning her photographs as soon as
possible so that any fast-moving objects which turned up could be
reported and followed up immediately. Only once the observing run
was over, and she was back in her ofce in Pasadena, California, did
T
h
e

C
e
n
t
a
u
r
s
31
she have time to return to the photographs and search for anything
moving slowly. Although the limited capability of the Palomar
Schmidt telescope meant that very faint objects were likely to avoid
detection, this time there was something recorded on the lm.
Although the Spacewatch team had been rst to make a report of
their detection, all the sets of positions were published together and
the new object was given the temporary designation 1992 AD. From
the preliminary orbit calculated using the January observations, it
was possible to begin a search for earlier detections. Soon a candidate
was found on a plate exposed a year earlier by the Shoemakers.
Assuming this sighting was indeed 1992 AD, another detection was
found by Beatrice Mueller on a 1989 image. Soon 1992 AD turned up on
photographs taken in 1982 and 1977. From all these observations a
denitive orbit could be calculated and the new object received the
minor planet number 5145. Now it needed a name. David Rabinowitz
was interested in moving away from the tradition of naming asteroids
after characters from Roman or Greek mythology, especially since he
felt that the Centaurs as a group were an unsavoury bunch. He
favoured naming outer planet asteroids after creatures from the cre-
ation myths of a number of different cultures. He suggested that the
new object be called Chaos. This would have been a very appropriate
name for an object in a planet-crossing, and so probably unstable
orbit. However, tradition prevailed and the object was eventually
named 5145 Pholus, who was Chirons brother.
Like that of Chiron, the orbit of Pholus is unstable over a timescale
of 10100 million years, but there the similarity ended. Soon after
Chiron had been discovered, astronomers had determined that its
surface is neutral in colour. This is to say that it reects all wave-
lengths equally and the light that comes back from it looks almost the
same as the sunlight which arrives there. Pholus was very different.
Within weeks of its discovery no less than three groups reported that
the new asteroid was astonishingly red. Although Mueller mentioned
it to Dave Rabinowitz, probably the rst to comment on this in print
was David Tholen from Hawaii. Tholen noted in an IAU Circular that
Pholus was the reddest asteroid he had ever observed. So unusual
were these colours, that Beatrice Mueller and her co-workers entitled
the paper describing their ndings Extraordinary colors of aster-
oidal object 1992 AD. Pholus, so much like Chiron in terms of its
orbit, seemed to be completely different physically. The rst sugges-
tion as to why Pholus was so red was that its surface was coated with a
B
e
y
o
n
d

P
l
u
t
o
32
layer, not necessarily all that deep, of carbon-bearing molecules
formed by the action of cosmic radiation on a surface that originally
comprised mainly simple ices. If this was true, then it would suggest
that Pholus surface was older than Chirons. Exactly how much older,
no-one could say.
This conclusion is of such importance to our story that it is worth
spending a few moments to see how it was reached. Astronomers use a
number of techniques to study asteroids and one of these is called
photometry. In photometry, the light arriving from an object of inter-
est is passed through a lter onto a detector (usually some sort of elec-
tronic device) which can record the amount of energy received. The
amount of energy from the source is then compared with that from a
star of known brightness which has been observed with the same
equipment under the same conditions. The ratio of the two values is
described in logarithmic units called magnitudes. Five magnitudes
corresponds to a brightness ratio of 100 and ten magnitudes to a factor
of 10 000. Photometry is used in all areas of astrophysics, but here we
need only concern ourselves with studies of asteroids.
If repeated photometric observations of the same object are made
over a period of a few hours or days, then it is possible to determine if
the object is varying in brightness. If this reveals a regular variation,
usually called a lightcurve, then the object is probably irregularly
shaped and is rotating, presenting different faces to the observer as it
turns. Lightcurves of small- to medium-sized asteroids often involve
changes of a few tenths of a magnitude over a period of a few hours.
For reasonably bright objects, lightcurve observations are quite easy
to do with even a relatively small telescope. Taking things a little
further, by making measurements in a number of different lters,
each of which pass only a narrow range of wavelengths, a sort of
crude spectral ngerprint can be obtained. This tells the astronomer
if the object reects more blue light than red, or red light than blue, or
if it reects all colours equally. The lter system at most observatories
uses ve colour lters called U (Ultraviolet), B (Blue), V (Visible
about yellow), R (Red) and I (Infrared). Subtracting the magnitude in
one lter from another gives what astronomers call a colour, for
example UB or BV. It had been known for some years that if the UB
and BV colours of asteroids are plotted on what is logically enough
called a two-colour diagram they are not scattered about randomly.
Instead, groups of objects cluster in specic regions of the diagram.
These different colours are ascribed to the presence of different
T
h
e

C
e
n
t
a
u
r
s
33
minerals, each of which reects light in a slightly different way. Thus
lter photometry is a popular tool of asteroid astronomers. The
reason for its popularity is that it is fairly quick and can be applied to
quite faint objects, so large numbers of asteroids can be observed and
classied. This helps pick out the unusual and interesting objects
from amongst the thousands of more-or-less ordinary ones. In this
case, Pholus stood out at once because of its very red VR and VI
colours.
Filter photometry is a good general tool, but provided that the
objects are bright enough a better way of nding out about the compo-
sition of astronomical objects is to take spectra. This involves using a
prism or a diffraction grating to spread the objects light out smoothly
over a range of wavelengths. Spectroscopy is used extensively in the
study of stars and galaxies since the spectra of hot objects often
contain narrow lines which can be used to identify specic chemical
elements and the physical conditions under which they exist.
Spectroscopy of asteroids is not so simple since any features tend to
come from molecules or minerals and are broader and much shal-
lower than atomic lines. This makes them harder to detect so spectro-
scopic observations have to be concentrated on the brighter asteroids.
Fortunately, both Chiron and Pholus are quite bright and both could
be observed spectroscopically.
The spectra of Chiron and Pholus in the region between about 0.4
and 1 microns, which roughly correspond to the UBVRI lters, are fea-
tureless. The optical spectra reveal nothing that might provide a clue
as to either objects composition. However, many simple molecules,
especially those containing carbon atoms, have spectral features in
the near infrared region of the spectrum. This region corresponds to
wavelengths from about 1 to 4 microns and for a variety of technical
reasons these wavelengths are much harder to observe than optical
wavelengths. Luckily, at about the time Pholus was discovered, a new
generation of infrared spectrographs was being put into service at
several of the worlds major observatories. One of these instruments,
the then new CGS4 spectrograph at the UK Infrared Telescope
(UKIRT) in Hawaii, was turned onto Pholus by staff astronomer
Gillian Wright in 1992. Ironically, Gillian Wrights main interests were
in extra-galactic astronomy, about as far as one can get from studying
solar system objects, but as the scientist responsible for the CGS4
spectrograph she was just looking for a suitable target for some tests.
She agreed to observe Pholus for a few minutes when it was suggested
B
e
y
o
n
d

P
l
u
t
o
34
as a possible target.

The observations were very successful. They

revealed that Pholus had infrared spectral features unique in the solar
system, notably an absorption in the spectrum around a wavelength of
2.25 microns. This feature could not be identied with any certainty,
but it was similar to features seen in chemical mixtures called tholins
produced during laboratory experiments attempting to duplicate con-
ditions in the early solar system.
The rst tholins were produced by taking asks of simple gases
which were expected to exist in the atmospheres of gas giant planets
and subjecting them to electrical discharges to simulate the effects of
lightning on the primitive atmospheres. The results of these experi-
ments were a mish-mash of carbon bearing molecules such as amino
acids and a red-brownish material that was quite stable and difcult
to destroy. This residue was named a tholin (from a Greek word
meaning dirty) and will be familiar to anyone who has tried to clean
up old glassware which has been used in organic chemistry laborato-
ries. The presence of tholin-like absorption features in the spectrum
of Pholus was conrmed when Dave Jewitt and Jane Luu of the
University of Hawaii repeated the observations a year later and,
thanks to technical improvements within CGS4 since the 1992 obser-
vations, obtained a much higher-quality spectrum. Jewitt and Luu
also observed Chiron with CGS4, but they found no sign of spectral
features similar to those in Pholus.
A number of other Centaurs have since been discovered, but many
of them are very faint and not much is known about them. The best
observed are 7066 Nessus (1993 HA
2
), 8504 Asbolus (1995 GO) and 10199
Chariklo (1997 CU
26
), all of which were discovered by the Spacewatch
team. Chariklo is quite bright and photometry of this object shows
that it is redder than Chiron, but less red than Pholus. Spectra from
both UKIRT and larger telescopes show that Chariklo seems to show
features due to water ice, but none of the Centaurs yet observed has
the deep 2.25 micron feature seen on Pholus. Paradoxically, ice is not
seen in most published spectra of 8504 Asbolus,

which looks other-

wise rather similar to Chariklo. However, there may be more to these
spectra than meets the eye. Spectra of Chiron from the mid 1990s
did not reveal the presence of ice there either, but more recent
T
h
e

C
e
n
t
a
u
r
s
35

The suggestion came from me, I happened to be visiting Hawaii and was sitting in on the
weekly UKIRT schedule planning meeting. It was a chance encounter that changed the
direction of my scientic research programme. JKD.

But see page 119.

observations have detected weak ice features. Chiron is known to
show comet-like activity and it seems that the ice has only become
visible since a period of activity a few years ago. This suggests that
the ice was always there, but that its spectral features may be easier
to detect under certain conditions than others. Small changes in the
structure or composition of their surfaces may explain why some
Centaurs do show evidence for ice and others do not. In all probability
they are all icy objects.
In 1998, Dale Cruikshank of the NASA Ames Research Center at
Moffett Field near San Francisco combined a number of observations
of Pholus with calculations of the spectra expected from various ices
and molecules to make a model of its surface. He believes that the
spectrum of Pholus can be explained by a combination of water ice,
a dark sooty material, some mineral dust and some ices of a simple
carbon-bearing molecule like alcohol. By adding these together in an
appropriate way, his group can match the observed spectrum of
Pholus quite well. Interestingly, although of course it is not a coinci-
B
e
y
o
n
d

P
l
u
t
o
36
Figure 2.4 The rst published infrared spectrum of the Centaur 5145 Pholus. Taken by
Gillian Wright from the UK Infrared Telescope UKIRT in 1992, it shows unusual spectral
features in the 2 micron region. The gap between 1.85 and 2.05 microns was not covered by
this observation. (Originally published in Icarus, vol. 102, p. 67. Courtesy John
Davies/Academic Press.)
dence, the materials chosen by Cruikshank for his model match pretty
well with our current understanding of what makes up the nucleus of
a comet.
Cruikshanks model offers a possible explanation of the dichotomy
between Pholus and Chiron. Perhaps they are essentially similar
objects, being giant comet nuclei composed of mostly primitive solar
system ices and some dust. They were probably formed about the
same time as the planets and have been stored in deep freeze in the
outer solar system since then. Recently, they were each moved
inwards by some subtle combination of gravitational forces. Perhaps
Pholus is still covered by an ancient crust formed when its ices, being
bombarded by cosmic rays (high-energy particles from beyond the
solar system) formed a layer of tholin-like material which helps to
preserve the ices frozen beneath the surface. Chiron, while basically
similar, is for some reason now subject to comet-like outbursts which
bring fresh material, probably ices, from its interior. Geysers of mate-
rial jetting out from cracks in the crust of Chiron may have fallen
back, or recondensed, coating the surface with a lighter, neutral crust.
This fresh material could be hiding any evidence for older material
that might be present. Either that or the outbursts have simply blown
away the ancient crust completely. Although plausible, if this explana-
tion is true it is not clear why Pholus has not experienced similar
activity.
Possibly either Chiron or Pholus was once nearer the Sun. If this is
T
h
e

C
e
n
t
a
u
r
s
37
Figure 2.5 The orbits of six
Centaurs compared with those of
Jupiter, Neptune and Pluto. At aphe-
lion the Centaurs are close to
Neptunes orbit, suggesting that they
might have originated in the trans-
Neptunian region. In this and the fol-
lowing gures the orbits of the Earth
and the other inner planets are not
shown as they would be almost indis-
tinguishable on this scale. (Chad
Trujillo.)
so then either of them may have had its surface drastically altered
before the random effects of the gravity of the inner planets ejected it
outwards again. This is by no means impossible. Calculations made
by Gerhart Hahn and Mark Bailey in 1990 suggest that there is a high
probability that Chiron has once been in an orbit more like that of a
short-period comet. If so, a close approach to the Sun could have
stirred up a burst of activity and removed its ancient crust leaving
behind a fresher and more active surface for us to observe today. On
the other hand, perhaps it was Pholus that ventured briey sunward
and experienced some drastic chemistry on its surface, and its red
colour today is just a bad case of cosmic sunburn. If the surface of
Pholus has been burned away, or chemically reprocessed, then
perhaps Chiron and the other Centaurs are more representative of
primitive solar system material and the idea that Pholus has an
ancient surface is just plain wrong. Unfortunately, the chaotic nature
of the Centaurs planet-crossing orbits makes it impossible to project
their positions far enough back in time to nd out.
The planet-crossing orbits of the Centaurs mean that they cannot
remain in their present locations for even a tiny fraction of the age of
the solar system. The fact that some, indeed an increasing number of
them, are being detected means that the Centaur population is being
continuously replenished. The discovery of the Centaurs lent cre-
dence to increasingly detailed suggestions that there was a reservoir
of objects beyond Neptune that could be diverted into Centaur-like
orbits and which would then evolve inwards to become short-period
comets.
B
e
y
o
n
d

P
l
u
t
o
38
The mystery
of the short-
period comets
By the early 1970s the rate at which short-period comets were being
discovered was increasing suspicions that there were too many to be
explained by objects captured directly from the Oort Cloud. One of the
people considering this problem was a professional physicist who had
started out with an amateur interest in astronomy. As an amateur
astronomer, Edgar Everhart discovered two comets, one in the
summer of 1964 and another in 1966. Perhaps it was this that sparked
his interest in the distribution of comets and led him to write his rst
scientic papers on the subject. His papers were well received and
marked a transition from the world of physics into that of astronomy.
Everhart assumed the directorship of the Chamberlin Observatory in
Denver, Colorado, and developed an interest in astrometry, the mea-
surement of the positions of objects in the sky. Everhart took pho-
tographs of faint comets and, in the basement of his home in the
Colorado mountains, he measured their positions using a machine he
had built himself.
Everhart became interested in orbital dynamics and started to
apply new methods of numerical integration to the study of how
comets orbits evolve under the inuence of the gravitational elds of
the planets. Although he had shown in 1972 that captures by Jupiter
were possible, it seemed to him that there were far too many short-
period comets to be explained by Jupiter captures alone. In 1977, fol-
lowing earlier work done by the Russian dynamacist
Kazimirchak-Polonskaya, he tried to solve the problem of the excess
of short-period comets. His method was to consider Oort Cloud
comets approaching the Sun on initial orbits that did not penetrate
the solar system as far as Jupiter, but which just skimmed the edge of
the planetary region in the vicinity of Neptune. The gravitational
C
H
A
P
T
E
R

3
39
effect on a comet during an encounter with Neptune is about 11 000
times less than the effect of an encounter with Jupiter, but Everhart
reasoned that Neptune does not have to do all the work. All that was
needed was for Neptune to change the comets orbit enough to move
its perihelion inwards, so that it has a chance of falling under the
gravitational inuence of Uranus, the next planet in. If this can be
done, then an encounter with Uranus may move the comet inwards
again until it falls under the gravitational control of Saturn. The
gravity of Saturn can then direct the comet into an orbit which passes
close enough to Jupiter for it to be captured into a short-period orbit.
Everhart simulated this process by examining the evolution of
thousands of cometary orbits on a computer using what astronomers
call Monte-Carlo methods. These are numerical techniques that take
their name from the famous casino town where small balls bounce
around roulette wheels and nish up in random slots around the edge.
The computing technology of the 1970s did not allow a detailed calcu-
lation of exactly how the orbits of each hypothetical comet would
evolve, so Everhart was forced to use a number of shortcuts. These
reduced the amount of computing time by a factor of 500, but even so
it required many hours to carry out the simulations. However,
although it worked in the sense that Everharts calculations showed
that short-period comets could be created in this way, the calculations
also showed that it was a very inefcient process. Many, many comets
were ejected from the solar system for each one that was captured.
The rst of the simulations considered the diversion of comets
towards Uranus by Neptune. These revealed that only 18 out of 12 230
simulated comets made the jump inwards. However, a second simula-
tion showed that once a comet had been steered into Uranus control,
its chance of going on toward Saturns sphere of inuence was 40/69.
Further numerical experiments showed that the likelihood of moving
from Saturns inuence to Jupiters was 229/500 and that, once there,
the likelihood of becoming a short-period comet in the Jupiter family
was 92/229. Since the nal likelihood of making a short-period comet
is reached by multiplying all of the above probabilities together, it
was clear that only one in several thousand comets would be captured,
the rest would be ejected from the solar system during one of the plan-
etary encounters along the way. The process was also very slow. A
typical comet that encroached Neptunes sphere of inuence would
survive over 40 million years before being sent inwards to Uranus.
Indeed, the capture of Oort Cloud comets by Neptune could only
B
e
y
o
n
d

P
l
u
t
o
40
resolve the problem of the excess of short-period comets if there was
a very massive inner Oort Cloud containing perhaps 700 times the
mass of the Earth. That such a huge mass of cometary material
should exist seemed unlikely. Another mechanism to supply the short-
period comets was needed.
A possible explanation came from Uruguaian astronomer Julio
Fernandez. While a student at the University of Montevideo,
Fernandez had become interested in the problems of the formation of
the solar system, how the planets had formed, how the direction of
planetary rotation was established and so on. However, in the late
1970s the political situation in his home country had become unstable.
Indeed, the mood throughout much of South America was generally
unfriendly towards science and other cultural activities. Feeling
uncomfortable with this environment, Fernandez moved to Spain. He
spent a year at the National Observatory at Madrid, where he contin-
ued to read and think about comets and their relevance to the forma-
tion of the solar system. Fernandez found the idea of a belt of comets
beyond Neptune put forward by Kuiper and Whipple exciting. Before
long he started to consider if such a belt could have anything to do
with the apparent excess of short-period comets. Since the rate at
which Oort Cloud comets could be captured into short-period orbits
was rather controversial, he attacked the problem from another point
of view. Using the results of people like Everhart, who had shown that
the most common fate of a comet from the Oort Cloud entering the
inner solar system was to be accelerated by planetary encounters
until it escaped from the solar system forever, Fernandez estimated
that for every comet nally captured into a short-period orbit, 600 or
more would be ejected into interstellar space. By combining the
number of known short-period comets and their expected lifetimes,
he was able to estimate the rate at which new ones had to be captured
to maintain a healthy comet population. From this, and assuming the
1:600 ratio of captures to losses was correct, he was then able to esti-
mate the rate at which Oort comets were being ejected into interstel-
lar space. The conclusion that he reached was that some 300 comets a
year were being lost. Such a high rate of wastage could not be main-
tained over the age of the solar system unless there were an improba-
bly large number of comets in the Oort Cloud to begin with.
Fernandez concluded that there must be an alternative source for the
short-period comets.
Having decided that short-period comets could not, in the main, be
T
h
e

m
y
s
t
e
r
y

o
f
t
h
e

s
h
o
r
t
-
p
e
r
i
o
d

c
o
m
e
t
s
41
captured from the Oort Cloud, Fernandez tried to decide if they could
diffuse inwards from a hypothetical trans-Neptunian comet belt. He
took as a starting point that there was about one Earth mass of comets
in a belt stretching from 35 to 50 AU. He assumed that this material
was distributed as bodies with sizes ranging from a small comet
nucleus up to something about the size of the Earths moon. He then
tried to estimate the effect of close encounters between objects in this
belt. His objective was to see if the gravitational effects of the largest
ones could perturb the orbits of the smaller ones into the planetary
region. His simulations of thousands of random encounters between
hypothetical trans-Neptunian comets soon indicated that a attened
disc of planetesimals orbiting just beyond Neptune could supply a
couple of comets per year into the region of the outer planets. Since
any objects sent inwards from this disc would start close to the plane
of the solar system, their further evolution would be quite rapid in
astronomical terms. Using Everharts estimates that there was a
roughly 50:50 chance that a comet crossing the orbit of one giant
planet would be sent inwards to encounter the next planet along,
Fernandez concluded that once a comet had moved into Neptunes
sphere of inuence it had about one chance in sixteen of evolving into
a short-period comet. The fate of the remainder was to be ejected back
outwards again. The important difference was that unlike the 1:600
estimate of captures direct from the Oort Cloud, a loss rate of sixteen
to one did not require an unrealistic number of Oort Cloud comets to
start with.
Having solved the problem to his satisfaction, Fernandez, who had
by now moved to a position at the Max Planck Institute for Aeronomy
in Lindau, Germany, wrote up his conclusions. He sent them for publi-
cation in the Monthly Notices of the Royal Astronomical Society, where
Edgeworths initial speculations had appeared three decades earlier.
No-one at the Royal Astronomical Society seems to have commented
on the omission of any references to Edgeworths work and
Fernandezs paper was published in 1980. At this point Fernandez felt
there was no point in continuing to work on the origin of the short-
period comets. He went back to working on issues associated with the
formation of the planets and the dynamics and origin of the Oort
Cloud. However, in closing his paper, he did pose the question of
whether objects in this hypothetical belt could be observed. Using an
estimate of the reectivity of Pluto as a starting point, he concluded
that objects in the comet belt might shine at about 17th to 18th magni-
B
e
y
o
n
d

P
l
u
t
o
42
tude. These would be difcult, although by no means impossible, to
detect with late 1970s technology. We now know that while Fernandez
was broadly correct in his conclusions about the existence of a trans-
Neptunian comet belt, he was wrong about this last point. As we will
see later, objects in the belt are much less reective than Pluto and so
they are much fainter than he thought.
A crucial advance in resolving the problem of the short-period
comets came in the late 1980s with a series of papers from a group of
theoretical astronomers based in Canada. A member of this group,
and someone who would go on to play an important role in under-
standing the dynamics of the outer solar system, was physicist
Martin Duncan. Born in London, England, in 1950 Duncan went, or
rather was taken, to Canada when he was 18 months old. He grew up
in Montreal. As a child, he always had an interest in how things
worked and admits to driving his parents crazy by constantly asking
typically small boy questions along the lines of, What would happen if
...?. At school he enjoyed mathematics and when he discovered
physics he thought it was a wonderful subject which would let him
study how things really work. This enthusiasm led him to take a
physics degree. Although he had never been an amateur astronomer
and still nds it terribly embarrassing when people ask him to point
out constellations in the night sky, it was during his degree that he
developed an interest in astrophysics.
Martin Duncan graduated during a time of great excitement in
astrophysics, with the discovery of exotic new types of objects such as
quasars, pulsars and black holes. Caught up in this excitement, he
went to the University of Texas at Austin to study for a PhD in general
relativity. He started to work on the theory of how gravitational col-
lapse forms black holes, but says he soon realised that solving this
problem was going to take more than the time allocated to getting a
single PhD. In the end, his thesis nished up being on the related
subject of the orbits of stars around black holes. Like many young
astronomers, he then had a fairly nomadic existence for a while. He
worked at Cornell University in up-state New York before going back
to Canada and spending ve years in Toronto as an assistant professor
at the university there. By then it was the mid 1980s and one of the hot
topics in astronomy was the suggestion that an unseen companion
star of the Sun, sometimes called Nemesis, might periodically inject
comets into the inner solar system. It was speculated that the result-
ing comet shower would cause large numbers of impacts on the Earth,
T
h
e

m
y
s
t
e
r
y

o
f
t
h
e

s
h
o
r
t
-
p
e
r
i
o
d

c
o
m
e
t
s
43
triggering mass extinctions like the one which removed most of the
dinosaurs about 65 million years ago. Becoming interested in this pos-
sibility, Duncan started reading more about this topic. He soon
realised that the injection of comets towards a very distant target,
such as the Sun or the Earth, was very similar to the problem of study-
ing the orbits of stars around a black hole. So, along with Scott
Tremaine, the director of the Canadian Institute of Theoretical
Astrophysics (CITA), a research establishment in Toronto, Duncan
put together a proposal to study cometary dynamics. Having been
awarded a grant to pursue this question, they hired a postdoctoral
research worker called Tom Quinn to help with the project. This
turned out to be a happy choice and led to a series of important papers
on the dynamics of comets.
Advances in computing power since Everharts work in the 1970s
now made it possible to do more detailed calculations of the path of
individual comets under the inuence of the gravitational forces of
the giant planets. In simulations described in their 1988 paper, The
origin of the short-period comets, Duncan, Quinn and Tremaine fol-
lowed the orbits of imaginary comets until they were either ejected
B
e
y
o
n
d

P
l
u
t
o
44
Figure 3.1 Martin Duncan, celestial
mechanician. (Terence Dickinson.)
from the solar system completely or became visible to equally imagi-
nary Earth-based astronomers. However, as Everhart had noted, the
diffusion of a comet into a short-period orbit can be a slow process. It
may take millions of orbits before the crucial interactions occur. So,
even with a new generation of faster computers, some approxima-
tions were still necessary to get the simulations done in a reasonable
time. One trick they used was to increase articially the mass of each
of the giant planets by a factor of up to 40. This increased the rate at
which orbital evolution occurred in their mathematical model.
However, even with this shortcut, the simulations still took several
months of computer time to complete.
First, they tried to duplicate Everharts conclusion that short-
period comets could be the result of the capture of Oort Cloud comets
with perihelia in a specially favourable region close to Jupiter
(46 AU). They created 5000 imaginary comets in their computers and
gave them a wide range of inclinations, allowing them to approach the
Sun from many directions. They then set the simulations running and
let the orbits evolve. As it turned out, they did not get the same result
as Everhart. They found just the opposite. Although comets entering
this region of the solar system could be captured, the population of
simulated cyber-comets which was produced did not look anything
like the actual short-period comets observed by real astronomers. In
particular, their simulations more-or-less preserved the inclinations
of the original comets and did not produce a population of short-
period comets close to the plane of the solar system. Next, they inves-
tigated Everharts other suggestion, that comets were captured by
Neptune and passed down from planet to planet until they reached
Jupiter. Once again, the resulting population did not look right; there
were too many comets in orbits with high inclinations compared with
the situation in real life. Their conclusion was that it was impossible
to form Jupiter-family comets in orbits close to the ecliptic plane if
they started with a population of Oort Cloud comets that could arrive
from all directions. The next step was inevitable, if the source was not
a sphere of comets, how about starting with a disc of material already
close to the ecliptic plane?
They set up their simulations again. This time they started with a
population of comets in orbits with inclinations of between zero and
eighteen degrees and on paths that already crossed the orbit of
Neptune. Once again they set the computers running and followed the
evolution of these cyber-comets as they either fell inwards towards
T
h
e

m
y
s
t
e
r
y

o
f
t
h
e

s
h
o
r
t
-
p
e
r
i
o
d

c
o
m
e
t
s
45
the Sun or were ejected into deep space. The result was dramatic. The
simulations produced a population of short-period comets very much
like the real thing. The small differences that still existed could easily
be accounted for by their assumption that all comets coming within
1.5 AU of the Sun are discovered (which may not be the case for real
comets) and by subtle effects caused by the need to increase the
masses of the giant planets used in the simulations in order to speed
things up.
Another conclusion of the study was that about 17% of comets that
start in Neptune-crossing orbits nish up as short-period comets. This
was three times more than Fernandezs estimate. While the disagree-
ment of a factor of three was not huge (some physicists joke that the
fundamental equation of astronomy is that 1 is approximately equal
to 10), it implied that the number of comets in the comet belt could be
correspondingly smaller. Only a fraction of an Earth mass of material
was now required. Furthermore, this material could be injected into
the Neptune-crossing orbits quite slowly, easily long enough for the
disc to have survived for the age of the solar system. For the moment,
it was not obvious exactly how material got from the comet belt into
the Neptune-crossing orbits, but once it got there it was clear that it
could supply the required number of short-period comets.
Duncan, Quinn and Tremaine went a few steps further in their cal-
culations. They showed that Chiron, which had been discovered by
Kowal a decade earlier, could well be a bright member of the parent
comet population which was in the process of undergoing just such
a diffusion inwards from what they called the Kuiper Belt. Finally,
they showed that for a reasonable size distribution, there ought to be
more than a thousand objects in this Kuiper Belt that would have a
magnitude of about 22, and so would be directly detectable with suit-
able ground-based telescopes. All that remained was to nd some.
B
e
y
o
n
d

P
l
u
t
o
46
Shooting in the
dark
Coincidentally, about the time that the existence of a trans-Neptunian
disc was being demonstrated mathematically, came the discovery of
an object that might be the rst member of such a structure. In
September 1992 two astronomers from the University of Hawaiis
Institute for Astronomy reported the detection of a faint, slow-moving
object orbiting the Sun beyond Pluto. The result took the community
of solar system astronomers by storm, but for Dave Jewitt and Jane
Luu it was not really a surprise. It was more a relief after ve years of
fruitless searching.
Dave Jewitt is a tall, thin Englishman with an acerbic wit and a
ne sense of the ridiculous. Gold-rimmed glasses and rapidly thin-
ning hair atop a wiry body create something of the illusion of a mad
scientist, but he is a shrewd and determined astronomer. Born in
Tottenham, North London, in 1958, he grew up in Eneld, another
London suburb. Jewitt discovered astronomy at the age of seven.
One evening in 1965 he was riding his bicycle home when he saw
large numbers of meteors, many of thembright enough to shine
through the haze of London street lights. Intrigued by this cosmic
rework display, he began to look more closely at the stars. Soon he
started to pick out some constellations and to learn his way around
the night sky. Not long after, his grandparents gave hima telescope as
a birthday present and with this he began to explore the craters on
the Moon, the rings of Saturn and the satellites of Jupiter. After
developing an interest in mathematics and physics at school, he
decided to study for a degree in astronomy at University College,
London.
While Jewitt was studying for his degree, spacecraft were return-
ing some of the rst good-quality images of the planets and revealing
C
H
A
P
T
E
R

4
47
landscapes shaped by hitherto unknown craters and volcanos.
Impressed by these discoveries, Jewitt thought he might like to pursue
a career as a planetary geologist. Since planetary exploration was
basically a NASA enterprise, he decided to do postgraduate work in
the United States. On the advice of a professor at University College,
London, he enquired about possible studentships at the University of
Arizona and the California Institute of Technology (Caltech), both of
which had strong planetary science research groups. He eventually
settled on Caltech because, he says, their application form was a lot
shorter and simpler to ll in.
Caltech is a large, famous and rich institution located on a campus
in the city of Pasadena, to the north of Los Angeles. It controls the
famous Mt Palomar observatory and also operates the Jet Propulsion
Laboratory (JPL), the nerve centre for numerous NASA space mis-
sions. While studying at Caltech, Jewitt worked on data from the two
Voyager spacecraft, which were returning stunning new data about
Jupiter and its satellites. He also began to use the telescopes on Mt
Palomar. During his trips to the mountain he learned about the reality
of doing astronomy for a living and (like many other neophyte
astronomers) had what he describes as many ugly experiences there.
However, in the process he learned how to do observational astronomy
in a very direct, hands-on manner.
In 1982, while still a graduate student at Caltech, Jewitt set about
recovering Halleys comet. Through his position at Caltech he had
access to the large telescopes on Mt Palomar and says he thought,
Gee, why not give it a try?. The search was carried out on various
occasions spread over several months, using a few hours here and
there borrowed from other astronomers who had been assigned the
200 inch (4.8 m) telescope for the night. The comet was eventually
found on the night of 16th October 1982, when Jewitt was observing
with Ed Danielson. It was not an easy observation; there was a bright
star very close to the predicted position of comet and the light from
the star was blinding the detector. Jewitt found an old razor blade in
his room and cracked it in half to make two knife edges. He put these
into a little ring which he mounted in the focal plane of the telescope
and used them to block off the bright star. The razor blade had obvi-
ously been used and had bits of stubble stuck on it. Jewitt and
Danielson called them Hubble hairs after the great Palomar
astronomer Edwin Hubble. The recovery of the comet, the rst time it
had been seen for over 70 years, was reported on the front page of the
B
e
y
o
n
d

P
l
u
t
o
48
Los Angeles Times, bringing good publicity to the observatory and
doing no harm to the career of Dave Jewitt.
Dave Jewitts thesis work included optical and infrared studies of
comets and he graduated with a PhD in 1983. That same year he left
California and moved to the East coast of the United States, becoming
an assistant professor at the Massachusetts Institute of Technology
(MIT) in Boston. Here he continued to work in planetary astronomy.
He made observations of comets and asteroids and published a series
of scientic papers, including some on the activity of distant comets
and the still-enigmatic Centaur Chiron. While at MIT he also began a
programme to search for other small distant bodies because he found
himself asking, Why does the outer solar system appear so empty?.
Having convinced himself that there must be something out there, he
set out to nd it.
Jewitt was joined inhis searchby a graduate student who had
arrived at MITby anevenmore convoluted route thanhis. LuuLe
Hang was borninVietnamand had come to the United States as a child
in1975, eeing Saigonwithher family as the NorthVietnamese Army
entered the capital. After settling inCalifornia, LuuLe Hang became
Jane Luuand set out to get anAmericaneducation. After learning to
speakEnglishand doing well inhighschool science classes, she
entered Stanford University and graduated witha Bachelors degree
inphysics. Althoughshe had planned to continue her career by study-
ing solid state physics, she was diverted into studying planetary
astronomy by chance. Working at the Jet PropulsionLaboratory as a
computer operator, where she had a routine job running computer pro-
grams connected withNASAs Deep Space Network, she was sur-
rounded by spectacular images of planets and moons whichadorned
the walls of the laboratory. Gradually these images seduced Jane Luu
into beginning a PhDinplanetary astronomy. She applied for, and was
awarded, a graduate student positionat MIT.
Dave Jewitt and Jane Luu did not work together directly at rst. As
part of her postgraduate studies, Luu had to complete two research
projects, one of several qualifying requirements which had to be com-
pleted before deciding on a subject for her PhD thesis. Her rst project
involved helping in a search for the optical counterparts of gamma
ray bursts, but at some point Jewitt suggested that a search for outer
solar system objects might make a good (if rather out of the ordinary)
research project. Jane Luu recalls asking in effect Hasnt this been
done already?. Jewitt replied No. She then asked Then why should
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
49
we do it? and was told Because if we dont, nobody will. She agreed
to give it a try.
They begantheir searchwithtraditional technology, using Schmidt
cameras and photographic plates inmuchthe same way as Clyde
Tombaughand Charles Kowal. The planwas that they would rst take
a photographof a regionof sky, thenreturnto that same patchof sky
after a short time to take another photograph. Thenthey would
compare the two images to see if anything had moved inthe interval
betweenthe two exposures. Althoughthey could expect to detect many
moving sources, it was possible for themto have a good idea of howthe
very-distant solar systemobjects they were seeking would be moving
across they sky. Since objects at the distance of Pluto take two to three
hundred years to go around the Sun, they only move across the sky at a
rate of about 1 degree per year. This is so slowthat it would be very
hard to detect were it not for the fact that the Earthis also moving
around the Sun. Since the Earthmoves comparatively quickly, it is
effectively overtaking the distant object. So, provided youlookina
directionthat is almost directly away fromthe Sun, towards what
astronomers call the oppositionpoint, thendistant objects appear to
move backwards onthe sky. This so called reex motionis about the
same for all distant solar systemobjects so there is a good chance that
any objects behaving inthis way are indeed inthe outer solar system.
In the rst half of 1987 Jewitt and Luu undertook several observing
B
e
y
o
n
d

P
l
u
t
o
50
Figure 4.1 Jane Luu and Dave Jewitt. The picture was taken in the control room of the
UKIRT telescope in 1994. (Jane Luu.)
campaigns using the Schmidt telescopes at two of Americas observa-
tories, the Kitt Peak National Observatory in Arizona and the Cerro
Tololo Inter-American Observatory in Chile. The Schmidt telescopes
could cover a large region of sky, each recording images covering an
area 5.2 degrees square in a single exposure. During each observing
run they photographed a series of regions along the ecliptic, generally
taking three images of each eld. Each image was exposed for about
one hour and the elds were rephotographed about a day apart, to give
slow-moving objects enough time to move signicantly between the
observations. Although only two such images were strictly necessary,
the third plate allowed the reality of any candidates to be checked, as
any suspicious object seen on a pair of plates should have a counter-
part with the same rate and direction of motion on the third.
The plates were examined during the week after the observing run.
This was done at the headquarters of the US National Observatories,
on Cherry Street in Tucson, Arizona, a few hours drive from the tele-
scopes on Kitt Peak. The searching was done using a blink comparator
installed in the basement of the building. According to Jewitt, the
comparator was serviced by a technician who could never be found,
but who would appear miraculously to x problems before he could
even be told about them and then vanish just as mysteriously after-
wards. Each pair of plates took about 8 hours to examine, and most of
them were searched independently by both Jewitt and Luu in order to
check that nothing had been missed. The process was hard work and
often excruciatingly dull. Jewitt recalls that it was difcult to main-
tain concentration for more than a couple of hours at a time. In the
1988 paper containing their results they describe the blinking process
as placing considerable strain on the eyes and it is worth considering
in a little detail exactly what was involved. Each Schmidt telescope
recorded its image onto a photographic plate 194 mm square, about
the same area as the open pages of a paperback book. A main belt
asteroid moves across about 1 degree of sky every four days, so during
a one hour exposure it moves about 30 to 40 arcseconds. This move-
ment corresponds to a trail less than half a millimetre long on one of
Jewitt and Luus plates. On a pair of plates taken a day apart the aster-
oid would appear as two such tiny trails, separated by about three
quarters of a centimetre, rather like the two horizontal parts of the
letters t in the word toot. More-distant objects, which were, of course,
what they were looking for, move more slowly across the sky. They
would appear as two streaks, each about a tenth of a millimetre long,
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
51
just 2 or 3 millimetres apart. Every square centimetre of each plate
had to be searched for evidence of such tiny marks, and every example
found had to be noted and then re-examined under a microscope to
check if it was real. Without exception, every candidate turned out to
be false, created by either tiny defects in the photographic emulsion of
the plate itself, or by specks of dust.
Luckily for Jewitt and Luu, and especially for their eyestrain, tech-
nology was moving on. By the late 1980s a new generation of elec-
tronic detectors called Charge Couple Devices (CCDs) were beginning
to challenge the photographic plate as the best way to make images of
the sky. Although when used with a suitable telescope, photographic
plates can cover a large area, even the best ones are relatively insensi-
tive. They convert only about 10% of the light which arrives into
exposed grains of the lms emulsion. In contrast, electronic detec-
tors can have very high efciency, as much as 90% or more in some
cases, but are rather restricted in size and so only cover small regions
of the sky in a single exposure. Obviously, searching the sky
efciently needs a detector which is both very sensitive (or goes deep
as astronomers like to say) and yet which can cover a large region of
the sky to increase the chance that there is actually something in the
image to detect. There is a trade-off to be made between sensitivity
and sky coverage. In 1987 it was not clear whether shallow and wide
was better than narrow but deep, so Jewitt decided to hedge his bets
and also try this new technology.
Today, CCDs form the backbone of almost every astronomical
detector in use at optical telescopes. They are made from semiconduc-
tor materials, usually silicon, which have been treated, or doped, in
such a way that when a photon of light falls onto the detector, it liber-
ates an electron from the semiconductor. These electrons, being nega-
tively charged, can be stored or moved about the device (which is often
referred to as a chip) by varying electrical potentials applied across it.
In operation, the CCD is mounted inside an instrument which is
exposed to the sky through a shutter, just like the lm in a normal
camera. However, when photons arrive at the focal plane of the
camera, they are not bound up in a chemical reaction with a lm
emulsion, but instead they release electrons into the thin wafer of
semiconductor material. At the start of each exposure the CCD is set
up with some xed voltages, so that any electrons liberated by the
incoming light stay put, and hot spots of electric charge build up
within the chip wherever light falls on it. Once the exposure has
B
e
y
o
n
d

P
l
u
t
o
52
been completed the shutter is closed. Then, by applying a series of
changing voltages across the chip, the electrons are read out in a
systematic manner which preserves a record of where on the chip
the charge came from. In essence, the chip is divided into a grid of
tiny squares called picture elements, or pixels, and the charge in
each pixel is recorded during the readout process. Once the readout
is complete, the voltages across the chip are reset to erase all evidence
of what happened during the previous exposure. Finally, the shutter
is opened again and the whole process is repeated.
One way to visualise the operation of a CCD camera being exposed
and then read out is to imagine a farmer trying to determine the way
rain is falling onto an open eld using a large number of buckets and a
team of labourers. First, our hypothetical farmer marks out a grid of
lines across the eld to make a number of squares. Next, the squares
are identied with letters and numbers, so that along one edge of the
eld the rows of squares have letters (A,B,C and so on) and up the edge
at right angles they have numbers. Each square can be identied by a
pair of letters and numbers such as A1, A2, B1, B2 and so on. The
farmer then sends a labourer with a covered bucket to stand in each
square and positions a measuring device, a scale or a measuring jug,
at corner A1. Each marked square in the eld corresponds to a CCD
pixel and the grid lines are usually referred to as rows and columns.
To start the measurement, analogous to opening the shutter in a CCD
camera, the lid of each bucket is opened. Raindrops (representing
photons, or rather the electrons generated in the silicon as the
photons arrive) are then collected for some xed time interval. To
stop the exposure our farmer blows a whistle and all the buckets are
covered over again. All that remains is to measure the amount of
water (photoelectrons) in each bucket and to plot that on a chart
representing the grid of squares across the eld.
It would be complete chaos if all the labourers tried to run over to
the farmer with their buckets, queue up to be measured and then run
back to their squares ready to repeat the measurement. So the farmer
has them all stand still and just measures the water in bucket A1. The
farmer notes the amount on a paper and pours the water away. Since
bucket A1 is now empty, the labourer in square A2 pours the contents
of bucket A2 into bucket A1 for measurement and recording. While
bucket A1 is being measured again, labourer A3 lls up the now empty
bucket A2 and waits to receive the contents of bucket A4. This process
continues until all the water in column A has been passed to the end
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
53
for measurement and all the buckets in this column are empty. At this
point, everyone in column B empties their bucket into the correspond-
ing bucket in column A and the process of measuring bucket A1
(which now contains the water originally in bucket B1), and passing
the water from bucket to bucket down column A for measurement is
repeated. While this second set of measuring and recording is going
on, all the buckets from column C are poured into those of the now
empty column B, the column C buckets are re-lled from column D
and so on across the eld. Once all the column A buckets have been
measured for a second time and are all empty again, they can be
relled once more from column B. The process of transferring water
across the eld, and then down one edge for measurement can be
repeated until every bucket is empty. Since the bucketfuls reach the
corner in a well-dened sequence (all the As, then all the Bs and so on)
the farmer can reconstruct the pattern of rainfall right across the
eld without ever moving from one corner. This process of pouring
water from bucket to bucket across the eld and down the edge is the
equivalent of reading out the CCD by varying the electrical voltages
across it to force the electrons to move about and then counting them
as they leave the chip.
Jewitt and Luus rst attempt with the new technology was made at
the relatively small 1.3 m telescope at the McGrawHill Observatory at
Kitt Peak, Arizona. The telescope was owned jointly by MIT and two
other universities and was equipped with a CCD camera developed by
George Ricker at Jewitts home institution of MIT. The camera was
known by the name MASCOT, the MIT Astronomical Spectrometer
Camera for Optical Telescopes. The MASCOT CCD was an array of
390 by 584 pixels, laughably small by todays standards, and for a
variety of reasons, not even all of this area could be used. Instead, the
useful area was restricted to a region only 242 by 276 pixels, giving a
eld of view of just 9 by 10 arcminutes on the sky. This was about one
thousandth of the area covered by a single plate from one of the
Schmidt telescopes used in the other part of the search programme.
The observing strategy which Jewitt and Luu adopted was to take
four consecutive exposures of each region of sky on one night, then
make two more of the same eld the following night to reveal any very
slow-moving objects. The exposure time for each image was about 20
minutes. If the exposures were any longer then even a slow-moving
object would begin to smear out across the image as it drifted across
the sky. This smearing would mean that the light from the target
B
e
y
o
n
d

P
l
u
t
o
54
would start to fall into adjacent pixels on the CCD and the signal to
noise ratio would no longer improve as the exposure continued. In
fact, the situation would soon start to get worse since the signal in the
rst pixel would no longer be improving while the continuing arrival
of unwanted background photons from the sky would steadily
increase the noise in it. Once the trailing effect becomes important,
doubling the exposure time just doubles the number of pixels into
which the light from the target is smeared and the signal in each pixel
no longer increases.
Over ve nights in April 1987 Jewitt and Luu observed 14 elds.
Since the computers on Kitt Peak itself were not then powerful
enough to process the images immediately, they took the data to
Cherry Street to examine the images there. This enabled them to take
advantage of the fact that not only are CCDs more efcient at detect-
ing light than photographic plates, but the fact that they are read out
electronically means that their images come out directly in digital
form. The digital data which describe the images can be either dis-
played immediately or further processed to reveal subtle structures
and objects that are not easily visible to the eye. One of the critical
steps in this process is at elding, which cancels out variations in
the sensitivity of different areas across the chip. Flat elding is
usually done using an image of a uniformly illuminated source, a
white screen or the inside of the telescope dome for example. The
resulting at-eld image maps out the pixel-to-pixel variations across
the chip and its inverse can be applied to all the subsequent science
images to atten them. The removal of false detections caused by
electronic noise in the measuring circuits and by cosmic rays passing
through the chip and liberating extra unwanted electrons is also
usually necessary. Once these steps have been taken, it is possible to
redisplay the image with different degrees of contrast or to try to
improve the ratio of signal to noise in the region of interest by adding
together groups of adjacent pixels. This latter process is called rebin-
ning, but of course, rebinning does not come free. When more pixels
are added together the ne detail in the resulting image is lost and the
picture becomes grainy and harder to interpret.
The switch to CCDs also highlights another advantage of electronic
detectors over photographic plates. Given adequate computing power,
the images produced from an electronic detector can be searched in a
number of different ways. The simplest of these is to subtract one
image from another electronically. All things being equal the xed
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
55
objects will vanish and anything which has moved between the expo-
sures will remain, appearing as a positive image from the rst frame
and a negative one from the second. Of course, in real life things are
never that simple so complete subtraction of the xed objects is
seldom possible. Small changes in atmospheric turbulence, or seeing
as it is known in the trade, and errors in the motion of the telescope as
it turns to counteract the Earths rotation, cause the images of stars to
smear out slightly. This means that the image quality is often different
from exposure to exposure. The images are also affected differently by
cosmic rays, with about 200 such hits being recorded in each MASCOT
frame. Despite these difculties, Jewitt and Luu found it fairly easy to
identify the relatively fast-moving main belt asteroids in the sub-
tracted images, but they soon found that subtracting images was not
the best way to search for faint, slow-moving objects.
Their preferred method was to blink images in much the same way
as was done using photographic plates. However, instead of having to
develop the plates, load and then carefully align them in a blink com-
parator before starting to search, modern image processing software
can do quite a lot to ease the strain of the blinking process. The soft-
ware can automatically match the overall brightness of pairs of
images and even align them before the blinking starts. Then, with a
couple of presses of a button, the images can be ashed alternately
onto a screen at whichever frequency the person doing the searching
feels most comfortable. This is usually about once per second. Not
only this but three, four or even more images can be blinked and the
image size and contrast can be varied quickly and easily to investigate
potentially interesting objects. It is here that the advantage of taking
multiple images of the same eld comes into play. Any ashing source
which does not appear in all the frames, and with a constant direction
and rate of motion, can be ignored. Real moving objects exhibit a con-
stant back and fro ashing motion which once seen and recognised is
very distinctive.
This is not to say that blinking images by computer is either easy
or quick. Any exposure deep enough to detect a new member of the
distant solar system will certainly be long enough to detect hundreds
of stars in our own galaxy and numerous faint galaxies beyond it.
There is also a good chance that the eld will contain a few main belt
asteroids as well. Despite the advantages of new technology, blinking
images is a very tedious business which demands a great deal of con-
centration. Typical CCD images contain hundreds of objects and to
B
e
y
o
n
d

P
l
u
t
o
56
pick out a faint one jumping backwards and forwards by a few mil-
limetres on a computer screen is a tricky task, especially when tired-
ness and boredom are setting in. Sometimes, just as when searching
photographic plates, changes in the seeing during the time between
the two images causes the faintest objects to vanish from one frame,
but not the other. If this happens, then chance alignments of faint
sources mimic the back and fro ashing of a moving object and these
can only be eliminated with careful examination of each image. The
key to success is having considerable patience and training the
eyebrain combination to recognise what looks right whilst ignoring
all the rest. Some people are better at this than others, so quite often
the best observing teams comprise pairs of people, one of whom oper-
ates the telescope and camera while the other blinks the images
relentlessly to see if anything can be found.
Although Jewitt and Luu both blinked their MASCOT images to
check that nothing had been missed, they did not nd any new distant
asteroids. This was perhaps not very surprising; the relatively small
size of the MASCOT CCD meant that only tiny areas of sky could be
searched at a time. Indeed, the area covered by their 70 CCD frames
covered only one third of a square degree, or less than one hundred
thousandth of the whole sky. However, the onward march of technol-
ogy soon led to an increase in size of a typical CCD detector and the
amount of sky that could be imaged in a single image grew steadily,
offering more and more chances of success. Dave Jewitt was not
deterred by his early, negative results.
In 1988 Dave Jewitt made another move. He left MIT to take a
faculty position at the Institute for Astronomy, or IfA, in Honolulu, on
the island of Oahu, Hawaii. The IfA owes its existence to the fact that
the summit of Mauna Kea, a huge dormant volcano on the nearby and
less cosmopolitan Big Island of Hawaii, is an excellent place to build
large telescopes. The Hawaiian island chain comprises several large,
and many small, islands stretching across a swath of the Pacic
Ocean. They were created as one of the Earths tectonic plates drifted
over a hot spot in the molten mantle beneath. Magma escaping
through the hot spot built a series of volcanoes, several of which rose
well above sea level. The extinct volcano of Haleakala rises some
3000 m over the island of Maui and has long been the site of a solar
observatory and a research facility of the US Department of Defense.
The story goes that astronomers working on Haleakala, including
ironically Gerard Kuiper, would glance across at the even higher
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
57
mountain of Mauna Kea on the Big Island and ask themselves, Why
dont we build an observatory there?. In 1964 a road, or more precisely
a dirt trail accessible by four-wheel-drive vehicles such as jeeps, was
cleared to the summit of Mauna Kea and testing of the mountain as a
possible place to put large telescopes commenced. The testing soon
showed that Mauna Kea was indeed an exceptionally ne observing
site. Kuiper described it as, Probably the best site in the world from
which to study the moon, the planets and stars. Three groups, Kuiper
from Arizona, a team from Harvard University and another from the
University of Hawaii all applied to NASA for funding to build a tele-
scope on Mauna Kea. In 1965 NASA awarded the University of Hawaii
a grant to develop a 2.24 m telescope on the summit. Soon after, the
state of Hawaii decided that the University of Hawaii should be
placed in charge of the development of Mauna Kea and so the IfA
was created near the Manoa Campus of the University, in Honolulu.
During the rst few years of the development of Mauna Kea, four
large telescopes were built there, including the 2.24 m optical tele-
scope which belonged entirely to the University of Hawaii and a
larger optical telescope, the 3.6 m CanadaFranceHawaii Telescope,
or CFHT. Joining these two were a NASA funded infrared telescope
devoted to observational support of space missions and the 3.8 m
United Kingdom Infrared Telescope (UKIRT). As the value of Mauna
Kea as an observatory site began to be recognised, more and more
observatories were established on the mountain and it is now one of
the most important astronomical observatories in the world. As land-
lord of Mauna Kea, the IfA collects a tithe of observing time from all
of the observatories on the mountain, so in addition to their own tele-
scopes, they have guaranteed access to 1015% of the time on all of the
other facilities. With so much telescope-time available to a relatively
small number of astronomers, the IfAs telescope-time allocation com-
mittee looked favourably on Dave Jewitts proposals. Soon he began
to receive allocations of observing time on the 2.24 m telescope for
various solar system projects. At rst, his programmes were not pri-
marily searches for distant objects, but as the capabilities of the tele-
scopes were improved, particularly in respect of bigger and better
CCD cameras, Jewitt began to think about doing some more surveys.
However, by then Dave Jewitt was not the only one searching for
trans-Neptunian objects. Between November 1988 and March 1989
Harold Levison, known to his friends as Hal, and dynamicist Martin
Duncan surveyed almost 5 square degrees of sky using the 1 mtele-
B
e
y
o
n
d

P
l
u
t
o
58
scope at the US Naval Observatory at Flagstaff, Arizona. The 2048
square pixel CCDthey used covered a square of 26 arcminutes a side
and with it they observed a total of 26 elds covering a total of 4.88
square degrees. They imaged each eld for 40 minutes at three sepa-
rate times spread over a couple of days. They did not search their
images by blinking them, but instead used computer software to iden-
tify all the objects in each image and then tried to link together
sources which did not appear at the same place in successive images.
This linking was done on the basis of pairs of sources having roughly
similar brightness and whose separation was consistent with motion
at the typical rate of a distant solar systemobject. The software did
indeed nd many such linked objects, between ve and 70 in each
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
59
Figure 4.2 The University of Hawaiis 2.24 m (88 inch) telescope used by Jewitt and Luu
during their search for the Kuiper Belt. One of the four-wheel-drive vehicles used to drive
up to the summit is parked outside. (John Davies.)
eld, all of which had to be examined individually. Every one of them
turned out to be false alarms, attributable to chance alignments of
cosmic ray hits, electronic noise in the CCDreadout and other defects.
So, despite their efforts, no distant objects were found, although
Duncan jokingly relates they nearly named a candidate after the
mythological character Ups, as in Whoops, whats that? Based on the
negative results of the search, Levison and Duncan attempted to place
limits on the number of Chiron-like objects and comets between about
25 and 42 AU from the Sun. For Chiron they compared the percentage
of the sky they surveyed (0.01%), the number of objects they found
(none) and the number of Chiron-type objects known at the time (one).
Based on the statistics of these very small numbers, they concluded
there were less than 7000 Chiron-sized objects in the outer solar
system. They also noted that this value, and their estimates of the pop-
ulations of still smaller objects, were very dependent on a number of
their assumptions and could not be taken too seriously.
Another search was carried out at the 2.7 m telescope of the
McDonald Observatory in Texas by Anita and Bill Cochran. Anita
came to the University of Texas as a graduate student in the late 1970s
and while there she met her future husband. Astronomer Bill Cochran
was a postdoctoral research worker in the university. Between
November 1990 and March 1993 they, mostly Anita, spent a total of 22
nights, including the Christmas eve of 1992, observing. Images were
taken with an 800 by 800 element CCD and were blinked to search for
moving objects. However, the search was dogged by poor weather and
poor seeing (which smears out faint objects over many CCD pixels and
makes them harder to detect above the background of the night sky)
and nothing was found. Realising that she could not compete with tele-
scopes at other sites, where better seeing made detecting faint objects
easier, and running out of time due to pressure of other projects,
Anita Cochran put her search programme aside for the moment and
moved on to other things. One of these was to be a project to search
with a telescope where seeing was not going to be a problem.
One other searchproject carriedout aroundthis time is worthyof
mention, eventhoughit was initiallycarriedout using images taken
for a quite different reason. InSeptember 1990, Piet Hut of the Institute
for AdvancedStudies at Princeton, NewJersey, was listening to a talk
about a proposedtelescope whichwouldsurveya large regionof sky.

B
e
y
o
n
d

P
l
u
t
o
60

This became the Sloan Digital Sky Survey.

The primarypurpose of the surveywas to make a huge census of
galaxies, but Hut suggestedthat it might also turnupsome rather
closer things. After the talk, Hut was approachedbyTonyTysonfrom
the nearbyBell Laboratories, who saidthat it was not necessaryto
wait for the surveyto be done. He andhis colleagues Raja
Guhathakurta andGaryBernsteinalreadyhadsome deepCCDimages
that might be suitable for just sucha search. Hut talkedwith
Guhathakurta about ways inwhichtheir data couldbe usedandthey
developedsome software to addtogether individual images inmoving
reference frames. Adding images inthis waysmears out the xedstars
andgalaxies, but adds together the signal frommoving objects. Of
course, this procedure onlyworks if youknowinadvance howfast the
target is moving. Luckily, since the motionof objects beyondNeptune
is dominatedbythe reexmotionof the Earthovertaking them, a good
guess at the correct rate of motionis possible evenwithout knowing
exactlyhowfar awaythe targets are. The searchalgorithms were then
testedonframes whichhadbeenmodiedto include a fake moving
object. The software seemedto workwell, so it was appliedto some real
images to see if anything couldbe found.
The method they settled upon was to use a number of their images
taken with relatively short integration times and combine them to
make a deep image in the reference frame of the xed stars. Then they
electronically subtracted these xed objects, and removed any defects
and cosmic ray hits from each of the original exposures. This left a
series of blank residual images. These were then added together
with successive offsets to allow for the motion of any hypothetical
faint objects travelling across the image in a certain direction with
a given speed. This rate of movement across the image is called the
objects apparent motion vector. The shifting and adding was then
repeated for a range of slightly different apparent motion vectors to
maximise the chance of nding faint moving objects.
Althoughthis methodwill workfor anypart of the sky, the best
chance of success will be for regions close to the ecliptic, where the
densityof potential targets is the highest. Tyson, Hut andcolleagues
triedtheir methodonsome images of a eld3.5 degrees fromthe eclip-
tic whichtheyhadtakenin1991. The eldwas observednine times
over a two-night periodandcovereda total area of 40 arcminutes a
side. Theyaddedthe images to cover a range of possible motions from
1 to 4 arcsec per hour. Nothing was found, eventhoughtheywouldhave
expectedto see objects as faint as 25thmagnitude. Theywere unlucky,
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
61
the methodwas a goodone andthe sensitivitytheyreachedwas high
enoughto have detectedsomething if it hadbeenthere. It just so hap-
penedthat the eldtheypickedwas emptyonthe nights theychose to
lookthere. Theygot, inthe words of Piet Hut, Close, but no cigar.
While Hal Levison, Anita Cochran and the others were trying and
failing, Dave Jewitt, who had been joined in Hawaii by Jane Luu in the
autumn of 1988, continued to search using the University of Hawaiis
(UH) 2.24 m telescope on Mauna Kea. Since it was from here that
success would eventually come, it is worth considering for a moment
exactly what they were up against. The summit of Mauna Kea is not
an easy place to work. At an altitude of 14 000 ft (4205 m), almost half
way to the height jet airliners y, the air is thin, with only 60% of the
oxygen found at sea level. It is often cold, and in the winter snow some-
times blankets the summit for days at a time. Almost nothing lives at
these high elevations. There is no vegetation on the reddish-brown
cinder cones and only a few very specialised insects survive by eating
other less hardy insects blown up the mountain to die. The thin and
often very dry air produces a number of interesting physiological
effects on the humans who choose to go there in search of clear skies.
These include headaches, itching eyes and nosebleeds brought on by
increased blood pressure as the heart tries to pump blood around the
body fast enough to ensure an adequate supply of oxygen to the brain.
There is also a little talked about tendency to run to the toilet rather
frequently as the kidneys try to thicken the blood by ejecting excess
water. To minimise the risk of altitude sickness, which is at best
unpleasant and at worst can be fatal, astronomers working on Mauna
Kea do not sleep at the summit. Instead they stay at a dormitory-cum-
cafeteria called Hale Pohaku (from the Hawaiian for house of stone)
situated at the 3000 m level, more than a kilometre below the tele-
scopes themselves. As an aid to acclimatisation, the astronomers
usually arrive at Hale Pohaku the night before their observing run.
This gives their bodies a chance to begin the adjustment process
before they go to work at the summit the next evening. Despite these
precautions, most regular users of Mauna Kea admit that their
thought processes at the summit are not as clear as they would like. If
pressed, most of them will admit to at least a few hours of almost
moronic thinking trying to solve a technical or mathematical problem
they could sort out in a matter of minutes in the thicker air at sea
level. It was in this environment that Dave Jewitt and Jane Luu con-
tinued their quest for the edge of the solar system.
B
e
y
o
n
d

P
l
u
t
o
62
Their technique changed little over the years. They tried to observe
on moonless nights, when the sky would be at its darkest and so faint
objects would not be lost in the glare of scattered moonlight. They
always observed in the spring and autumn. These periods were chosen
since it is the time of year when the ecliptic plane, their best hunting
ground, was well separated from the Milky Way. Both the ecliptic and
the Milky Way, the plane of our Galaxy, trace out great circles across
the sky, but these two circles are not the same. This is because the
solar system is tipped relative to the plane of the Galaxy. The great
circles cross in two regions, in the northern sky at the border of the
constellations Gemini and Taurus, quite close to the famous Crab
Nebula, and in the south in the SagittariusOphiuchus region. Since
the faint band of light which we call the Milky Way is actually the
combined starlight of most of the stars in our own Galaxy, it is a very
crowded region of sky, already bedecked with a myriad of faint stars.
It is hardly the place to look for a faint point of light moving against a
background of many other points of light. Far better to look away
from the Milky Way, in the constellations of spring and autumn. Here
the sky is less crowded and any interlopers are easier to spot and less
likely to be masked by the light from background stars.
For Jewitt and Luu, each night began with an early dinner at the
halfway house of Hale Pohaku and a chat with some of the other
astronomers or perhaps some last-minute preparations for the observ-
ing. After dinner came the 20 minute drive, by four-wheel-drive
vehicle, up the twisting and bumpy dirt road that connects Hale
Pohaku to the summit. The UH 2.24 m telescope is on the third oor of
a traditional white observatory building, cylindrical in shape and
topped by a dome. From certain angles, the outline of the dome is
broken by an unusual hammerhead extension which is the parking
spot for a crane used to move heavy equipment about inside. The
building has laboratory space and workshops on the lower oors. The
control room is reached by taking a lift part way, passing through a
small workshop which smells of oil and machine tools, and then
climbing a short ight of spiral stairs.
Described by one of the staff as the highest ofce in the world the
control room itself is banana shaped, long, narrow and curving to
match the inside wall of the cylindrical building. At one end, a door
opens onto a catwalk around the outside of the building. From the
catwalk all the other observatories and a superb view of the Mauna
Kea sunset are visible. Ranged along the inner edge of the control
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
63
room are consoles for the telescope operator, who is responsible for
pointing the telescope and generally keeping things running
smoothly. Next to them are the computers of the observers station,
from where the scientic instruments are controlled. At the other end
of the room, an electric kettle and jar of instant coffee hint at a long
observing night. An internal door gives access to the observing oor
and to the telescope itself.
The telescope is a stumpy tube, rather like a barrel, held inthe arms
of a massive metal forkwhichtowers over the observing oor. The tele-
scope is painted orange-brown; it is said that the paint was blended
especially to matchthe colour of the dust and rocks of the summit in
order to bring good luckto the observatory. The base of the telescope,
where the scientic instrument for the nights observing is tted, is
about four metres above the oor. It canbe reached by a hydraulic plat-
formwhichcanbe raised to provide a working area around the tele-
scope. The CCDcamera operates best whenit is cold, so before
observing canbeginit must be topped up withliquid nitrogencoolant.
Once everything was ready and the dome had been opened, Jewitt,
Luu and their telescope operator settled down for the nights observ-
ing. Night after night they took a series of images, each exposed for 15
minutes, and returned to the same eld three times over the following
two hours. They worked as a team. One of them concentrated on the
image processing and blinking, the other controlled the camera and
kept track of the observing. As the images came in, they were dis-
played on a computer dedicated to data reduction located in the tele-
scope control room. Jewitt and Luu blinked as they went along
although not everything could be done at once, despite having what
for the time were quite powerful computers. Blinking the images
demanded so much from the computers that it could only be done
during the exposures when the CCD was inactive. Once the process
of reading out the CCD was about to begin, the blinking had to be
stopped. If it was not, the overloaded computing systems found it
impossible to cope and crashed.
The initial observing runs did not result in a rapid success. The
pair detected hundreds of main belt asteroids, but they ignored these
in order to concentrate on their more distant goal. However, by the
early 1990s CCD technology was moving steadily forward and larger
and larger chips were becoming available almost every year. The
2.24 m telescope soon moved from a 385 by 576 CCD to one 800 pixels
square. This seemed huge at the time, but was in fact only a step to
B
e
y
o
n
d

P
l
u
t
o
64
still better things. Soon, the CCD chips had increased in size to 1024 by
1024 pixel arrays. This steady increase in detector size meant that it
was easy for Jewitt to convince himself that success was only a matter
of time and to maintain his enthusiasm for the search. Jewitt helped
persuade Gerry Luppino, who had been a graduate student at the
Massachusetts Institute of Technology and who was known as an
excellent builder of CCD cameras, to join the IfA. By 1992 the 2.24 m
telescope was equipped with a camera containing a chip with 2048 by
2048 pixels able to cover an area of sky about 7.5 arcminutes square.
With it Jewitt and Luu continued to search for their elusive goal.
Sometimes they worked quietly, sometimes the control room pounded
to the sound of Jewitts CD collection, very heavy metal music from
groups with names like Cannibal Corpse, Mortician and Napalm
Death. Jane Luu was not a big fan of Jewitts musical tastes. When it
got too much for either her or the telescope operator, trapped in the
same room night after night, they switched to something from Luus
more restrained classical music collection. Sometimes they forsook
the CD player for Mauna Kea gossip and long rambling quasi-philo-
sophical conversations about nothing in particular.
The Jewitt and Luu combination was something of an attraction of
opposites. From the outside it seemed incredible that the outspoken
and amboyant Jewitt and the reserved and cultured Luu would be
compatible, but they got along well and were a very good observing
team. They were both determined to get the most of every minute of
telescope time. According to one of the telescope operators who
worked with him, Jewitt would even plan his observing to minimise
the number of times the telescopes protective dome had to be moved.
Every move was a couple of precious minutes lost. The pair made a
very productive partnership, publishing papers on the physical prop-
erties of cometary nuclei, on the statistics of the Earth-approaching
asteroid population and on the activity of the Centaur Chiron. Jane
Luu wrote a thesis entitled An observational investigation of the
comet-asteroid connection, and left the IfA for a postdoctoral position
at Harvard University in 1990. She came back to Hawaii from time to
time to continue the search.
The long-awaited breakthrough came around midnight on 30th
August 1992. It was the second night of a ve-night observing run.
Jewitt and Luu were taking the third of a sequence of four images of a
eld in the constellation of Pisces, which they called simply ECL 123.
Dave Jewitt was blinking the rst two images from the set. Suddenly
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
65
he said, Jane, come and have a look at this. There was something
starlike moving slightly between frames as they ashed alternately
onto the screen. This seemed to be exactly what they were looking for,
or was it? With only two images it was a bit early to be sure. Perhaps it
was a false alarm caused by cosmic ray hits or some other defect.
However, they were excited enough to measure the rate of motion,
which turned out to be 2.3 arcseconds per hour. This would place the
object about 60 AU from the Sun. When the third and then the fourth
image conrmed that they had indeed found a faint slow-moving
object, they were dumbstruck. Jewitt later described it as a sort of
mental collision between seeing the object, which looked just right,
and the experience of years of fruitless searching. Jane Luu recalls
I think we jumped up and down for a little while. The only question
was whether it was indeed very distant or something much closer.
There was a chance that it might be a small, near-Earth asteroid
moving almost parallel to the Earth and which happened to appear
almost stationary on the sky at that particular moment.
The only way to check if they had nally found what they wanted
was to make more observations. Anxiously they re-observed the
object several times over the rest of the night. They were looking for
any change in its rate of motion. Such a change would reveal that the
object was so close that as the Earth turned the different relative posi-
tions of the object and telescope would alter the observing geometry.
No such changes were found. Before breakfast they called Brian
Marsden at the Minor Planet Center to let him know they had found
something that looked interesting, but they agreed to keep it quiet
while they collected more observations. Over the remaining three
nights they re-observed the object repeatedly. They checked their mea-
surements again and again looking for anything which they might
have overlooked and which might destroy their conclusion that the
object was far beyond Neptune. They measured its brightness, posi-
tion and motion as accurately as possible and then they sent their data
to the Minor Planet Center. On 14th September Marsden announced
the discovery via IAU circular number 5611. He gave the new asteroid
the prosaic designation 1992 QB
1
.

B
e
y
o
n
d

P
l
u
t
o
66

The subscript 1 indicates that in that fortnight more than 24 objects had been reported
and the Minor Planet Center had had to use the letter B more than once. In the following
years, a dramatic increase in the number of asteroid sightings from automatic tele-
scopes would mean that much larger subscripts would be needed.
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
67
Figure 4.3 Some images of 1992 QB
1
taken on the night of the discovery. A main belt
asteroid appears in the top three images, moving from right to left. The motion of 1992 QB
1
is very much slower. (Dave Jewitt.)
From the limited number of positions available he could not
conrm exactly where the new object was; in fact, Marsden could not
even be sure which way it was going around the Sun. The rst posi-
tions could be tted with an orbit at about 41 AU going in the same
direction as the rest of the planets or with one some 15 AU further out
and going the other way. However, he did note that some solutions
were compatible with membership of the supposed Kuiper Belt,
although he qualied this remark by saying that the object could also
be a comet in a near-parabolic orbit. Further observations, over a
longer arc, would be required to solve this problem.
Luckily, these further observations were not long coming. As the
Moon waned and the skies became dark again, 1992 QB
1
was observed
from the Anglo-Australian Observatory at Siding Springs, New South
Wales, on 21st September. It was seen again from the European
Southern Observatory in Chile a few days later. Using these positions,
Marsden calculated another orbit which placed 1992 QB
1
a few AU
further out, but which was otherwise quite similar to his rst attempt.
However, the issue of the nature of the new object was still not closed.
Its positions could also be tted by other more-eccentric orbits.
Marsden thought it possible that 1992 QB
1
was a Centaur like Pholus
and Chiron, but that it had been discovered at the farthest point in its
orbit, rather than when close to the Sun. If this was true then there
was a chance that a pre-discovery image might appear on an old photo-
graph taken when 1992 QB
1
was closer to the Sun and so brighter.
Marsden, who enjoys entering into lighhearted wagers, found what
looked like a suitable candidate from 1930 and he made a bet with Dave
Jewitt that observations over the next few months would reveal that
the object was a Centaur.
Not just the orbit of 1992 QB
1
was uncertain, almost nothing was
known about its physical properties either. From the limited observa-
tions available, Jewitt and Luu deduced that their discovery was quite
red. This was unusual for an asteroid and hinted that 1992 QB
1
might
resemble the Centaur Pholus, discovered earlier the same year. They
searched for, but could not nd, any trace of gas and dust, which more-
or-less ruled out it being a large and very distant active comet. Even
the size of the new object was uncertain; without knowing details of
how reective its surface was, the discoverers could only make an edu-
cated guess at its diameter. Assuming that the reectivity was about
4%, a typical number for a dark asteroid, they estimated that the new
body was quite small, about 250 km in diameter. This would make it
one ninth the size of the planet Pluto.
B
e
y
o
n
d

P
l
u
t
o
68
S
h
o
o
t
i
n
g

i
n

t
h
e

d
a
r
k
69
Figure 4.4 The IAU circular announcing the discovery of 1992 QB
1
. The circular contains
some brief discovery details, a list of the positions reported by Jewitt and Luu and some
comments from Brian Marsden. These are followed by approximate orbital elements and
predicted positions for the next few weeks based on these elements. (CBAT/Brian
Marsden.)
Unless 1992 QB
1
was indeed a Centaur then its faintness meant that
there was little hope that images of it would be found on older photo-
graphic plates. It was clear that it would take some time to dene the
objects orbit with any great precision. However, six months later,
Jewitt and Luu discovered a second object, designated 1993 FW, which
was of similar brightness and which was moving in the same general
way. When this was announced there was little doubt that the exis-
tence of a trans-Neptunian population had nally been conrmed.
Marsden, being nothing if not a man of his word, paid up on his bet
with Dave Jewitt at a scientic conference in Flagstaff, Arizona, in the
summer of 1994. By then, four more distant asteroids had been found
and Marsden was giving a review of the discoveries and of the likely
orbits of the the new objects. Almost at the conclusion of his talk he
called on Dave Jewitt and solemnly handed him ve banknotes in
front of the assembled audience. It was all done in such good humour
that few of the people in the theatre realised how much money was
changing hands. Each of the ve banknotes was a $100 bill!

Brian Marsden got some of his money back two months later when, at a different
meeting, Dave Jewitt bought him a very expensive lunch.
B
e
y
o
n
d

P
l
u
t
o
70
Deeper
and deeper
With the discovery of 1992 QB
1
and 1993 FW, Dave Jewitt and Jane Luu
had opened the door on a new eld of solar system research. Not since
Piazzi discovered the rst main belt asteroid in 1801 had an entirely
new class of solar system object been discovered. In contemplating
the implications of the discovery, Dave Jewitt mused that,
Discovering the Kuiper Belt is like waking up one morning and
nding that your house is ten times as big as you had thought it was.
However, nding two objects was a beginning, not an end. To nd out
just how large the Kuiper Belt really was would require the discovery
of many more trans-Neptunian objects. Jewitt did not rest on his
laurels; indeed, his success seemed to spur him on to greater efforts.
Together with Jane Luu, who was still working in California, he con-
tinued searching. During an observing run in September 1993 they
found another two faint, slow-moving objects. These were soon desig-
nated 1993 RO and 1993 RP. Brian Marsden announced the discoveries
via IAU Circulars 5865 and 5867 on the 18th and 20th of September,
respectively. He noted that his initial calculations suggested that the
new objects appeared to be closer to the Sun than the rst two discov-
eries, but that with so few observations, the orbits could not be estab-
lished with any certainty. Marsden noted that the two objects could be
in circular orbits about 39 AU from the Sun, more eccentric orbits at
various distances or even in a variety of retrograde orbits. The obser-
vations could also be tted by parabolic trajectories as if the objects
were long-period comets entering the solar system from the Oort
Cloud. Marsdens dilemma was compounded a few days later when
another group entered the fray, reporting the discovery of two more
faint, slow-moving objects.
One of the members of this group, who would go on to make other
C
H
A
P
T
E
R

5
71
contributions to this new eld, was a young scientist named Alan
Fitzsimmons. Like many other astronomers, Fitzsimmons had
become interested in the heavens as a child and his hobby gradually
became his career. As he recalls it, he was about twelve or thirteen
when he visited a friend who was aching to show off the telescope he
had just received as a present. Fitzsimmons was talked into having a
look at the Moon and, impressed despite himself, he decided to save
enough money to buy a telescope of his own. Like Dave Jewitt and lots
of other young people, he slowly became hooked on astronomy. At the
time Fitzsimmons career goals did not include science; he saw his
future in business and he planned to study physics and computing at
university with this in mind. It was only a chance conversation with
the brother of a friend which alerted him to the possibility of study-
ing astronomy at university. After completing a rst degree at Sussex,
he went on to do a PhD at Leicester University. His project involved
studies of Halleys comet and brought him into contact with some of
the UKs few solar system astronomers.
Leaving Leicester, and the solar system community, he moved to
Queens University in Belfast, Northern Ireland. Here he began to do
research on the chemical composition of stars and eventually
accepted a permanent post in the Physics department. While he was at
Belfast, he saw the IAU circular announcing the discovery of 1992 QB
1
and thought to himself We can do this. After talking with Professor
Iwan Williams from Queen Mary and Westeld College in London, an
observing crony from his Halleys comet days, he contacted Dave
Jewitt. Together they applied for telescope time to make a further
search. Williams and Fitzsimmons were successful at the rst
attempt. They were awarded a week of observing time on the 2.5 m
Isaac Newton Telescope (INT) on La Palma, in the Canary Islands.
The observing run took place in late September 1993 using a 2048
pixel square CCD camera. Dave Jewitt was unavailable, he was
observing in Hawaii, so Fitzsimmons and Williams were joined on the
run by Belfast student Donal OCeallaigh. They used the same observ-
ing technique as Jewitt and Luu, taking and then blinking 30 minute
exposures aimed at a swath of positions along the ecliptic plane. The
project got off to a good start with a discovery on the second night of
the run. The object was faint and it required the taking of a few more
images to conrm its reality. Once these extra frames had been taken
there was no doubt that there was something there. The objects posi-
tion was measured and transmitted to the Minor Planet Center, which
B
e
y
o
n
d

P
l
u
t
o
72
designated it 1993 SB. The following night they found another object.
The second discovery, 1993 SC, was so much brighter than the rst one
that Fitzsimmons later described it as being, blindingly obvious in
the blinked frames.
Things were going so well that they were not very surprised to nd
another bright candidate two days later. Amazed by their good
fortune, and thinking that this business was much easier than they
had expected, they carefully measured the positions of the new object
and dutifully reported them to the Minor Planet Center. Brian
Marsden replied by remarking that, based on the positions, their
second and third objects seemed to be one and the same. It only took a
minute to conrm this and realise that, carried away by their enthusi-
asm, they had made a mildly embarrassing mistake. Somehow they
had made an error when calculating the location of the ecliptic plane
on the sky. Instead of continuing on into unexplored territory, they
had doubled back and accidentally rediscovered 1993 SC. The odd
thing was that, due to its slow motion, the object had hardly moved
from where it had been two nights earlier. In all the excitement not one
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
Figure 5.1 The Isaac Newton Telescope building on the island of La Palma in the Canary
Islands. Unusually for a modern observatory, ofce and workshops are located around the
telescope dome. The fth and sixth trans-Neptunian objects were discovered from here by
Iwan Williams and Alan Fitzsimmons. (John Davies.)
of the three observers had noticed that object number three was in a
stareld that they had searched before and which, in hindsight,
looked rather familiar. As with 1993 RO and 1993 RP, the initial obser-
vations suggested that the new objects were located in the general
region of Neptune, but, once again, with only a few days of observa-
tions accurate orbits could not be determined.
Alan Fitzsimmons subsequently led several other observing runs
to La Palma, discovering a number of other trans-Neptunian objects.
Meanwhile, Jewitt and Luu continued to make discoveries from both
Mauna Kea and from a small telescope in Chile. Other astronomers
were not slow to realise that there was much to be done at the new
frontier of the solar system and soon other groups joined in the hunt.
Over the next few years the rate of discovery increased steadily as
more and more people developed an interest in these obscure solar
system objects. Paradoxically, it was not long before the problem was
not so much discovering new objects, but keeping track of the ones
that had already been found.
It was soon apparent that a proper understanding of the Kuiper
Belt would require two lines of attack. On the one hand detailed physi-
cal studies were needed to probe the sizes, shapes and chemical com-
positions of individual objects. The rst step along this road was for
the orbits of individual objects to be determined accurately. Once this
was done their positions could be predicted well enough to allow them
to be studied with large telescopes. On the other hand, there was inter-
est in a more general understanding of the population of the trans-
Neptunian region as a whole. Accurate orbits for a large number of
objects would eventually be needed to provide theoretical
astronomers with information about the dynamics of the trans-
Neptunian region. The theoreticians wanted real orbits which they
could compare with the phantom objects predicted by their computer
models.
However, since an object at a great distance from the Sun moves
only slowly across the sky, observing its motion over a few days or
weeks is not sufcient to determine its orbit precisely. Orbits can be
described by numbers called orbital elements, and once these are
known accurately the objects position can be projected well into the
future. A typical set of orbital elements comprises six numbers, each
usually quoted to ve, six or even seven decimal places. It is the job of
people like Brian Marsden and his colleagues at the Minor Planet
Center to work them out. To do this, Marsden needed follow-up obser-
B
e
y
o
n
d

P
l
u
t
o
74
vations of each object a few times during the year of its discovery and
again from time to time over the next few years. This long period of
observation was demanded by the objects great distances from the
Sun. Even ve years of observations would cover only 2% of a typical
orbit. The problem was that with visual (V) magnitudes of only about
24, about sixty million times fainter than the faintest star visible to
the naked eye, any kind of further observations required observing
time on moderately large telescopes. Without this vital astrometric
follow-up there was a real risk that some of these objects might be lost
as uncertainties in their expected position built up to the point that it
would be impossible ever to nd them again. The situation was eerily
reminiscent of that almost two centuries earlier. Then too, asteroids
were sometimes discovered one year only to be lost again later. In the
nineteenth century, when orbital calculations were done on paper by a
small band of mathematicians, the problem was one of a lack of com-
puting power. In the late twentieth century there was plenty of com-
puting power; the problem was a lack of observing time on big
telescopes.
Big telescopes around the world are the front line of modern
astronomy and naturally large numbers of astronomers want to use
them. Since there is never enough telescope time to go around (a
typical national telescope receives three to four times more observing
requests than it can accommodate) astronomers compete with each
other through a process called peer review. What this rather
grandiose title actually means is that once or twice a year
astronomers write proposals asking for a certain amount of telescope
time. These proposals, which are usually a few pages long, are then
reviewed to decide which projects get an allocation of telescope time
in the coming months. This competition is judged by other
astronomers who try to be unbiased, but who generally have their own
ideas on what is important for the advancement of science and what is
not. There are trends, or fashions, in science just as there are in other
things, and at any given time there will be areas of astronomy which
seem to be producing the most interesting results. It is these elds
which may seem to offer the best chance of critical discoveries that
will produce a real scientic breakthrough, perhaps even a Nobel
prize or two. Conrming the orbits of a few distant asteroids, however
interesting they might be to a small number of planetary scientists, is
not seen by many astronomers as being on the cutting edge of modern
astrophysics. So, with the best will in the world, allocation committees
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
75
were sometimes reluctant to award time for apparently routine
requests for follow-up measurements of newly discovered Kuiper Belt
objects. There may well have been a feeling that, Some-one else must
be doing it, but in fact, a lot of the time, nobody was.
Marsden and a few other astronomers were acutely aware of this
problem. As early as 22nd September 1993, when he was still wrestling
with deciding exactly where in the solar system to put the four objects
discovered that autumn, Marsden had written that There is rapidly
developing a severe problem of securing adequate astrometric follow-
up which is absolutely essential for any understanding of this excit-
ing development in the outer system. People like Jewitt, Luu,
Fitzsimmons and their collaborators were conducting a determined
rearguard action of astrometric follow-up, but they were often ham-
pered by bad weather or baulky equipment. Soon it seemed that
objects were being lost almost as quickly as they were discovered.
Anita Cochran, who had been scooped in her own Kuiper Belt search,
contributed to follow-up observations. Mark Kidger from Spain, David
Tholen and Brett Gladman were amongst the few other dedicated
souls willing to use the limited telescope time they were allocated for
the essential, but routine astrometric follow-up. The magnitude, if the
reader will forgive the expression, of this problem was brought home
by the appearance of astrometric observations from a new observa-
tory site unfamiliar to most professional astronomers. The little
known Cloudcroft observatory was in fact the home of an amateur
astronomer named Warren Offutt who was carrying out a programme
of astrometric follow-up of trans-Neptunian objects for fun!
Even amongst a group of people as notoriously difcult to pigeon-
hole as astronomers, Offutt stands out as something out of the ordi-
nary. He was born in a suburb of New York City in 1928. This was two
years before Pluto, let alone its brethren in the Kuiper Belt, had been
discovered. He then had a successful career in industry before retiring
in 1990. Unlike the professional astronomers exploring the Kuiper
Belt, who are beholden to their employer or some other organisation
to fund their research, Warren Offutt could afford to have his own
observatory and then use it as he chose. Located near the village of
Cloudcroft, New Mexico, not far from the professionally operated
Apache Point observatory, Warren Offutts 60 cm telescope sits on a
knoll a few dozen metres from his house. It is inside a traditional
masonry building capped by a dome which contains both the telescope
and its control system. Offutt did not build this telescope himself.
B
e
y
o
n
d

P
l
u
t
o
76
Rather, as bets a man with decades of experience in the manufactur-
ing industry, he drew up a detailed specication of what he wanted
and negotiated with someone else to build it for him.
Although he had once started a project to photograph all the plane-
tary nebulae visible from Cloudcroft, Offutt abandoned this idea half
way through when he discovered that someone had already done it.
Luckily, as is often the way in astronomy, a new project came along at
just the right moment. At about 8 oclock one morning in 1995 local
astronomer Alan Hale called and asked him to verify the position of a
new comet he had discovered. Using the positions, estimated direction
and rate of motion supplied by Hale, Offutt looked for the supposed
comet the next night. After a short time it was clear that the comet
was not where he expected to nd it. Where had it gone? Not to be
beaten, Offutt scanned along an arc of sky covering the likely track
and detected the comet on his third photograph, taking what may
well have been the rst post-discovery image of Comet HaleBopp.
Although it was soon clear that the comet was beyond Jupiter, and so
must be unusually bright, determining how close it was going to come
to the Sun turned out to be problematical at rst. Many observers
took photographs to help nd out just what was going on and Offutt
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
Figure 5.2 Warren Offutt inside his private observatory in New Mexico. (Warren Offutt.)
supplied about 50 measurements of HaleBopps position as part of
this effort. Looking back, he says he found this a very satisfying expe-
rience, as he enjoys projects which can be described quantitatively, i.e.
in terms of precise numbers. The HaleBopp experience proved inter-
esting enough to make him continue with astrometric work, and
having set off along this course the challenge of looking for fainter
and fainter objects was irresistible. In a manner that seems almost
inevitable to him, but absolutely astonishing to many other people, he
found himself making astrometric observations of objects at the very
edge of the solar system.
For Warren Offutt a typical observing session starts off with a list
of specic objectives for the night. He usually aims to detect and
measure the positions of a selection of asteroids fainter than about
magnitude 18. Offutt restricts himself to such faint objects since if he
does not, he often comes across other asteroids by chance. When this
happens his quest for completeness makes him feel obliged to follow
them up as well. Before long the number of objects he is trying to
track proliferates so quickly that he says it becomes a chore, not a
pleasure, to observe them. If the observing goes according to plan, he
is usually nished by about 1 a.m. If not he sticks it out all night to get
his target list completed. Hopefully, things will go well as Offutt
admits that he is not physically up to observing the whole night for
several nights in a row. However, like many astronomers a third his
age, he still feels bad about stopping observing when the weather is
good and says that when he does, his conscience bothers him a little.
To him, the tracking down of objects that would have been almost
undetectable to even a professional astronomer using the best equip-
ment in the world only a few decades ago is an interesting project with
which to test the quality of his telescope and his personal observing
skills.
When looking for these very faint objects Offutt usually makes a
number of 2030 minute exposures. During each exposure his tele-
scope tracks the stars using an automatic guider which, after some
years of tuning, will keep it on target for hours at a time. Over the
course of a night he makes eight or so images and then adds two or
three of them together to bring out the faintest objects. Adding in
more than three images does not help. The motion of even a very
distant asteroid is enough to move it noticeably after an hour or so and
adding more frames does not improve the objects detectability unless
the frames are shifted to allow for the objects motion before being
B
e
y
o
n
d

P
l
u
t
o
78
added. Shifting and adding images in the reference frame of the aster-
oid is a common trick for professional astronomers who have large
computers, professionally written image analysis packages and full-
time software engineers to help them, but Offutt only does it in very
special cases. As he puts it, The extra processing increases the labour
greatly and I have to do everything here, including vacuuming the
oor. When not vacuuming, Offutt was instrumental in providing
recovery observations for a number of distant objects and as a tribute
to his dedication to this most arcane of observing tasks, Minor Planet
number 7639 was named Offutt in his honour.
The success of groups discovering objects using what might be
described as medium-sized telescopes encouraged others to try their
luck with bigger instruments. The objective of these searches was to
nd smaller, and so fainter, trans-Neptunian objects to establish how
the number of small objects compared with the number of large ones.
This information, which is called a size distribution, can help reveal
the history of the Kuiper Belt. It may tell astronomers if the objects in
the trans-Neptunian region are still growing or if, like the main belt
asteroids, they are now slowly grinding themselves down into dust. A
few groups decided that instead of searching large areas of sky for a
few bright objects, they would concentrate on going very deep over
small areas to see how many really faint objects turned up. Depending
on the details of the size distribution, this can actually be quite a good
way of nding things since, for most populations, there are a lot of
small objects for every large one. So, provided the search goes deep
enough, the smaller area searched will be balanced by a larger
number of potential discoveries. Since to go deep requires staring for
long periods at a single location rather than covering a wide area,
these surveys are often called pencil-beam surveys as they are akin
to sticking a pencil through the sky to see what can be found.
Several pencil-beam searches have been conducted by Brett
Gladman and colleagues using both the 3.6 m CanadaFranceHawaii
Telescope (CFHT) on Mauna Kea and the 200 inch (4.8 m) Hale
Telescope at Mt Palomar. Gladman, who at the time was based at the
Canadian Institute for Theoretical Astrophysics (CITA) in Toronto,
was interested in the problem because he realised that it was possible
to push the detection of faint objects to the limit by using large tele-
scopes, sensitive detectors and novel observational techniques.
Gladman felt that the searches being carried out by Jewitt, Luu and
others in the mid 1990s were too conned in the range of magnitudes
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
79
which they were nding to reveal much about the size distribution of
the trans-Neptunian population. So, since wide-eld cameras able to
scan huge areas in search of the rare, large objects were not yet avail-
able, he set about nding the smallest objects possible.
One search was done from Mt Palomar in September and October
1997. A total of ve nights of observing time with the 200 inch Hale
telescope were committed to just three elds, each of a square of a
little less than 10 arcminutes a side. Since the survey was aimed at the
faintest detectable objects, the searching was not done by blinking
pairs of images. Gladman used a two-stage process which began by
adding together all the frames of each eld using a variety of motion
vectors expected for typical trans-Neptunian objects. These combined
frames were then searched for any evidence of faint point-like sources
amidst the forest of trailed star and galaxy images smeared out by the
shifting process. Simple visual inspection of the images was not
enough. Gladman and his friends found that the best way to search
was to blink a series of different images of the same region which had
been combined using different apparent motion vectors. When they
did this they found that real objects showed a very distinctive pattern,
they got brighter as the frame with the most nearly correct apparent
motion vector was approached and then faded away afterwards. Once
something had been found, or its presence suspected, the process was
repeated using a ner range of motion vectors. Eventually, the correct
rate of motion was identied and the moving object became a well-
dened point source in the image.
The strategy they chose was to observe close to opposition and to
pick elds which were relatively empty. Choosing empty elds min-
imised the risk that the objects they were seeking would be lost in the
glare frombright stars. They also chose elds which they knewcon-
tained a previously identied trans-Neptunian object. Although the
presence of the known object could not be counted for statistical pur-
poses, it could be used to check their data reduction technique.
Measurements of the known objects position could also be used to
update the astrometric database of observations for orbital calcula-
tions. The rst eld they chose was one near an object designated
1996 RR
20
. It turned out to be a happy choice as blinking the images at
the telescope immediately revealed a newbright candidate, subse-
quently designated 1997 RT
5
. Coincidentally Alan Fitzsimmons inde-
pendently discovered this object the same night. Fitzsimmons was at
the Isaac Newton Telescope in La Palma conducting a programme of
B
e
y
o
n
d

P
l
u
t
o
80
astrometric follow-up of known objects and was observing 1996 RR
20
for this very purpose. Gladman located two other faint objects when
the shifting and adding technique was applied to his data. One of
these two was subsequently recovered in October. The recovery
observations were made by Elanor Helin and Dave Rabinowitz, the
discover of Pholus, who had by nowleft Spacewatch and moved to
California. The other object, detected with a magnitude of about 25.6,
was not recovered and has nowbeen lost.
A second eld was targeted on the position of a relatively bright
object designated 1996 TO
66
. As well as this fairly bright object, the
images also revealed a much fainter unknown one with an R (red)
magnitude of about 25.8. It was designated 1997 RL
13
and was at the
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
81
Figure 5.3 Four faint candidates are visible in this image taken as part of Brett
Gladmans search for small trans-Neptunian objects. (Brett Gladman.)
time the faintest object ever given a minor planet designation. The
object is probably only 40 kilometres in diameter and was about
44.5 AU from the Sun at the time of its discovery. It has not been seen
again. The third eld, which was observed under rather poor condi-
tions in October, did not reveal any new objects. Observations from
the CanadaFranceHawaii Telescope in April did not go as deep, but
they did reveal one new object on just a single night. Normally, a
single night observation would be of little use, but the same eld
was observed by Alan Fitzsimmons group a few days later.
Fitzsimmons observations allowed the object to be identied on
several other images and permitted the designation 1997 GA
45
to be
assigned. Since then Gladman has moved to the Observatory of Nice
on the south coast of France and has found a number of still fainter
objects. An observing run on the CFHT in February 1999 led to seven-
teen new objects being catalogued. One of these was designated 1999
DG
8
which, with an R magnitude of 26.5 and a distance of 62 AU,
became both the faintest and the most distant trans-Neptunian object
to receive a minor planet designation. Another observing run in
January 2000, also at the CFHT, revealed two more very distant objects
(2000 AC
255
AC and 2000 AF
255
) both about 53 AU from the Sun. In addi-
tion to continuing his searches, Gladman has also been trying hard to
recover some of the very distant objects discovered by himself and
others.
Similar deep surveys were carried out by Jane Luu and Dave Jewitt
over a range of nights between 1994 and 1996 and by Caltech
astronomers E. I. Chiang and Micheal (Mike) Brown in 1997. These
observations were made with the 10 m Keck telescope in Hawaii. Each
group detected a few faint objects, but their orbits remain very uncer-
tain. This is hardly surprising since, almost by denition, these deep
surveys nd mostly very faint objects. Astrometric follow up of these
faint objects is almost impossible without large amounts of time on
big telescopes, and this is seldom available. It was only a series of
coincidences which led to some of the objects in Gladmans survey
being observed by other telescopes and such coincidences tend to be
few and far between. Also returning to the hunt for faint trans-
Neptunians was Gary Bernstein, who with Piet Hut had missed out on
nding the rst Kuiper Belt object as early as 1991. Bernstein was
using data from the same telescope as last time, the 4 m Victor M.
Blanco telescope at Cerro Tololo in Chile, but this time instead of the
measly 1024 by 1024 pixel CCD available in 1991 he was using a new
B
e
y
o
n
d

P
l
u
t
o
82
instrument called the BTC, or Big Throughput Camera. The BTC com-
bines four 2048 by 2048 pixel chips to make a CCD array with a stagger-
ing 16 million pixels, able to cover half a square degree of sky. These
images were searched using the technique they had used unsuccess-
fully in 1991, combining several 10 minute exposures of the same eld
using a range of different motion vectors. This time, however, the
larger collecting area of the BTC did the trick; eight new objects were
found during three nights of observing.
As more objects were discovered and their orbits conrmed, it
became obvious that there was more to the trans-Neptunian region
than met the eye. As early as 1994, in IAU Circular 5983, Brian
Marsden had pointed out that while the rst two objects discovered
were in more-or-less circular orbits beyond Neptune, several of the
autumn 1993 discoveries were not. The other objects seemed to be in
elliptical orbits which came close to, or crossed, the orbit of Neptune.
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
83
Figure 5.4 1999 DG
8
, the most distant trans-Neptunian object yet observed. Several hours
of exposures from the 3.6 m CanadaFranceHawaii telescope have been reconstructed to
mimic a brief exposure from a telescope 50 m in diameter. This makes the object stand out
without being buried by the trailed images of stars in the eld. (Brett Gladman.)
Knowing that close approaches to a planet can change the orbit of a
comet or asteroid and send it inwards towards the Sun, or eject it from
the solar system entirely, Marsden suggested that the orbits of the
autumn 1993 objects were rather special. Specically, he suggested
they were in orbits which took them around the Sun with a period that
was an exact fraction of Neptunes orbit. Under certain circum-
stances this would protect them from being unceremoniously moved
into another, far less stable orbit. The resemblance of the situation of
these objects to that of Pluto, which also crosses the orbit of Neptune,
led Dave Jewitt to refer to then as Plutinos or little Plutos. The reali-
sation that Plutos orbit was not unique also had another side effect; it
eventually started a debate about the planetary status of Pluto itself.
Having dened one class of Kuiper Belt objects as Plutinos, it
seemed inevitable that names would be invented for other orbital
groupings that were being identied. Apart from the Plutinos, most of
the remaining objects were in roughly circular orbits about 4245 AU
from the Sun. This is distant enough that they are almost unaffected
B
e
y
o
n
d

P
l
u
t
o
84
Figure 5.5 The orbits of some of the better known Plutinos. The orbits of Neptune and
Pluto are included, but Pluto is indistinguishable from the others. (Chad Trujillo.)
by the gravitational forces from the rest of the planets and their orbits
are fairly stable. The rst object found in such an orbit was 1992 QB
1
so, following an astronomical tradition of naming objects after the
rst member of each class discovered (e.g. T Tauri stars and BL Lac
objects), Brian Marsden suggested that these objects might be
referred to as Cubewanos (Q-B-1-ohs). This name did not nd much
favour amongst the rest of the astronomical community. Instead, and
apparently spontaneously, the distant objects began to be called
Classical Kuiper Belt objects, recognising that they were in orbits
which most resembled the predictions made by Kuiper and Edgeworth
in the 1940s and 1950s.
An object in a rather different type of orbit was discovered in
October 1996 by Luu, Jewitt and his graduate students Chad Trujillo
and Jun Chen. Designated 1996 TL
66
, it was discovered on 9th October
at the University of Hawaiis 2.24 m telescope on Mauna Kea using a
new, mammoth CCD camera containing no less than eight, 2048 by
4096 pixel, detectors. Built by Gerry Luppino and known as the 8K
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
85
Figure 5.6 The orbits of some members of the Classical Kuiper Belt. These near circular
orbits do not cross that of Neptune. (Chad Trujillo.)
array the new camera had an effective size of 8192 by 8192 pixels and a
eld of view about 18 arcminutes square. Since most of the three
dozen or so Kuiper Belt objects which had been discovered up until
then were faint, the new detector was being used to a search a rela-
tively large area of sky in the hope of detecting a few brighter objects
that might be good targets for detailed physical studies. The huge size
of the new array made blinking the entire images impossible, and so
they were searched using a computer program developed especially
for the task as part of Trujillos PhD project.
Trujillos Moving Object Detection Software searched sets of
images of a single region of sky which had been taken in fairly quick
succession. After the usual processing steps to at-eld the images
and determine the sky background, all the objects in each image were
identied automatically. Cosmic ray hits and other defects were
removed and then the images were aligned electronically. All the sta-
tionary objects were then discounted. Finally, the software looked for
objects which were starlike, about the same brightness in each image
and seemed to be moving with a constant velocity in the range
expected for a distant solar system object. The coordinates of any can-
didate objects which passed these criteria were listed and circles were
then drawn automatically on the images, highlighting the regions
containing potential candidates. Once the few areas of interest were
so marked, the images were then blinked to check the reality of each
candidate. This nal step was required since the human eye is actu-
ally much better at discerning a real moving object from chance align-
ments of noise than even very advanced computer software. Trujillo
estimates that using his software to highlight only the most promising
detections for human inspection cuts down the amount of time
required to search the plates by what he calls an order of magnitude,
making a task that would otherwise have taken several hours per
frame possible in just a few minutes.
At 21st magnitude, 1996 TL
66
was the brightest trans-Neptunian
object found to date. While initial attempts to determine its orbit sug-
gested that it might be a Plutino, follow-up observations in December
1996 by Carl Hergenrother of University of Arizona were not consis-
tent with such an orbit. Brian Marsden deduced that either there was
something wrong with one of the sets of observations or the orbit was
rather unusual. Marsden asked Warren Offutt if the Cloudcroft
Observatory could obtain further observations to clarify the situa-
tion. Offutt once again stepped into the breach left by the telescope
B
e
y
o
n
d

P
l
u
t
o
86
time-allocation committees. After taking a series of observations on a
particularly good night in January 1997, and devoting some tender
loving care to drag the faint signal of the asteroid out of the noise in
his image, he was able to measure its position well enough for
Marsden to solve conclusively the mystery of 1996 TL
66
. The entire
series of observations showed that 1996 TL
66
was close to the perihe-
lion of a highly eccentric orbit with a period of 788 years. This orbit
takes 1996 TL
66
over 130 AU from the Sun, well beyond the connes of
the classical Kuiper Belt. The orbits of 1996 TL
66
, and a few similar
objects discovered a few years later, are highly elliptical. These objects
are in or near the classical Kuiper Belt when closest to the Sun, but
spend the rest of their time very much further away. Their huge ellip-
tical orbits take them hundreds of astronomical units from the Sun.
They are known as scattered disc objects.
Chad Trujillo is not the only person using software, rather than
manual blinking, to search for distant asteroids. The Spacewatch
project, which in 1992 discovered the second known Centaur, 5145
Pholus, has long been interested in using computer power, rather than
brain power, to search its large and ever-expanding datasets.
Spacewatch was the brainchild of asteroid astronomer Tom Gehrels,
who had long promoted the idea of a telescope dedicated to studying
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
87
Figure 5.7 Left to right. Dave Jewitt, Chad Trujillo and Scott Sheppard examine an image
taken from the 12K CCD camera on the CanadaFranceHawaii telescope. Trujillos
moving object software marks each candidate Kuiper Belt object by drawing a circle
around it. (John Davies.)
88
F
i
g
u
r
e

5
.
8
T
h
e

o
r
b
i
t
s

o
f
f
o
u
r

s
c
a
t
t
e
r
e
d

d
i
s
c

o
b
j
e
c
t
s
.
T
h
e

P
l
u
t
i
n
o

r
e
g
i
o
n

a
n
d

t
h
e

C
l
a
s
s
i
c
a
l

K
u
i
p
e
r

B
e
l
t

a
r
e

s
h
o
w
n

t
o
g
e
t
h
e
r

w
i
t
h

t
h
e

o
r
b
i
t
s
o
f
t
h
e

p
l
a
n
e
t
s

J
u
p
i
t
e
r

a
n
d

N
e
p
t
u
n
e
.
(
D
a
v
e

J
e
w
i
t
t
.
)

the small objects in the solar system. Although the idea for such a tele-
scope had been around since about 1969, the small number of
astronomers interested in asteroids meant that it was not until about
1980 that the project began to take shape.
In 1981 a historic 0.9 m telescope became available to Gehrels and
his small group. It was the original Steward Observatory telescope,
rst erected on the university campus in Tucson, Arizona in 1921 and
moved to the clearer and darker skies of Kitt Peak in 1962. Spacewatch
acquired the telescope and by 1983 was operating it with a 320 by 512
pixel CCD camera. In line with other observatories, Spacewatch
rapidly acquired larger CCDs, moving to a 2048 square device in 1989.
Since it is a dedicated asteroid search telescope, and has to cover large
areas of sky repeatedly rather than going very deep on small areas,
Spacewatch operates differently to most of the other groups we have
so far encountered. Instead of pointing at a selected region of the sky
and then turning to counteract the Earths rotation, the Spacewatch
telescope does not rotate with the sky. Instead, it points at a xed posi-
tion and allows the image of the sky to move across the detector.
Under normal circumstances this would result in the star images
smearing out, or trailing, and would be quite unacceptable. However,
the Spacewatch CCD is aligned on the sky so that individual star
images move parallel to the edge of the chip. In effect the stars drift
along a single row during the exposure. By adjusting the rate at which
the charge in each pixel is transferred to the next column during the
readout process, the reading of the array is synchronized with the
rate of drift of each star across the chip. When the accumulated
charge reaches the column at the edge of the chip it is transferred
down the end column and out of the array in the usual manner. This
readout technique is called drift scanning. It is very efcient since the
CCD is constantly exposed to light (there is no shutter opening and
closing) and there is no dead time between exposures while a shutter
is closed and an array is read out. In the context of the farmer trying
to measure the rainfall across the eld described in Chapter 4, the
buckets are being emptied into each other across the eld at exactly
the same rate that small clouds are drifting overhead. So when the
buckets reach the end of the eld, each contains water from just a
single small part of the cloud directly above it. Drift scanning for
asteroids at Spacewatch began in earnest in 1984.
Spacewatchs automatic Moving Object Detection Program
(MODP) came into use in 1985. Its development required about eight
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
89
B
e
y
o
n
d

P
l
u
t
o
90
Figure 5.9 The Spacewatch telescope on Kitt Peak, in Arizona. The astronomer is Jim
Scotti. (Lori Stiles, University of Arizona.)
man-years of work, mostly by Dave Rabinowitz and James Jim
Scotti. Scotti was born in 1960 in a small coastal town called Bandon,
in Oregon, where he lived for 8 years. As a boy he was captivated by
the sight of astronauts walking on the Moon and became interested in
space travel. As the Apollo programme came to an end, his interest
expanded to encompass not just the Moon, but also the rest of the
solar system. He bought his rst telescope with money he had earned
by babysitting and his interest in astronomy blossomed. He went to
the University of Arizona and was awarded a degree in astronomy in
1982. The very next week he started working for Spacewatch full time,
having previously worked there as an undergraduate. He now spends
about six nights per month observing at Spacewatch.
During normal observing the Spacewatch telescope is used to scan
a strip of sky about half a degree wide and seven degrees long three
times each night. The MODP software records the positions and
brightnesses of all the objects detected during each scan. When
observing is nished, the software compares the three lists and looks
for something which appears to be moving in a consistent way from
scan to scan. Usually many hundreds of such candidates are recorded.
These candidates must then be examined by the observer to decide
which are real and should be followed further and which are just
chance alignments of noise or bad pixels and can be ignored. This
checking takes several hours of the daytime, allowing the observer
only a limited time to eat, sleep and prepare for the following nights
observing. Of course, the number of false detections can be reduced
by making the softwares selection criteria more rigorous, but experi-
ence has shown that if this done some real detections are also missed.
Although only applicable to very rapidly moving objects, and so not
relevant to our story, Spacewatch also has software which searches for
streaks caused by objects moving so quickly that they trail out during
a single exposure. Using this fast-moving-object software Spacewatch
has discovered a number of asteroids which pass very close to the
Earth.
Although originally associated more with discovering objects
passing close to the Earth, the Spacewatch project always had in mind
the search for more-distant objects. Tom Gehrels describes the overall
goals as discovering small objects throughout the solar system in
order to study their statistical and dynamical properties. The discov-
ery of Pholus in 1992 drew attention to Spacewatchs capability to
detect more-distant objects and this was emphasised when Jim Scotti
D
e
e
p
e
r

a
n
d

d
e
e
p
e
r
91
discovered three more Centaurs. Although Dave Rabinowitz has
moved on, Scotti has remained at Spacewatch. He still has his boyhood
enthusiasm for astronomy, but these days he sometimes regrets that
during a typical observing session there is never enough time to stay
outside and really enjoy his view of the sky.
Since it rst went into operation the Spacewatch telescope has
scanned a very large area of sky. The huge amount of data it has taken
has recently been searched to see if any very-distant objects might
lurk there. New software was written to nd slow-moving objects
undetected by the original MODP software. On 17th March 1999 this
new software had its rst success, nding an object which had
remained hidden in the data since September 1995. This object, 1995
SM
55
, was re-located from the ground in 1999 and its orbit is now quite
well dened. Soon after, six more objects, all with visual magnitudes
between 20.6 and 21.5, were located in the data archive. Although some
of these have not been observed again, others (including objects desig-
nated 1998 SN
165
, 1998 SM
165
and 1998 VG
44
) have all since been re-
located and now have secure orbits.
The Spacewatch team continues its search and by mid 2000 it had
discovered two new and interesting objects which do not t into any of
the three previously dened groups. These two objects (1996 GQ
21
and
1999 TD
10
) have orbits which sweep across those of some outer planets
like the Centaurs, but which are not conned to the main planetary
region. At their furthest from the Sun these objects travel deep into
space, reaching distances of over 140 astronomical units away. These
large perihelion distances are more like those of the scattered disc
objects. Not long after these discoveries the Minor Planet Center
merged its lists of Centaurs and scattered disc objects. This decision
reects the fact that there are probably no fundamental differences
between the two classes of objects, other than that they have been
gravitationally scattered in different directions.
B
e
y
o
n
d

P
l
u
t
o
92
Sorting out the
dynamics
As early as spring 1994 Brian Marsden had remarked that some of the
newly discovered trans-Neptunian objects had orbits that were
similar to that of Pluto, but what is it that makes Plutos orbit special?
The general details of the ninth planets 246-year orbit had been estab-
lished within a year of Tombaughs discovery and two things made it
stand out from the other planets. Plutos orbit is the most highly
inclined, and the least circular, of all the planets. In fact, the orbit is so
eccentric that it crosses the orbit of Neptune. How can Pluto survive
in such an orbit?
Ancient philosophers believed that the Earth was the centre of the
Universe and that the planets moved in circles. This geocentric model
survived until Nicolas Copernicus proposed a heliocentric (Sun
centred) model of the solar system in the early sixteenth century.
Although he was correct in his belief that the Earth went around the
Sun, Copernicus erroneously assumed that planetary motions must
be based on circular forms. The last vestiges of heavenly perfection
were removed the following century when Johannes Kepler showed
that the orbits of the planets are not circular, but elliptical. However,
the amount by which the orbits of the major planets differ from a
circle is generally quite small, so much so that if a diagram of the
solar system is drawn to scale on a normal-sized piece of paper, the
average eye would be hard put to detect the elliptical nature of most
planetary orbits. Mathematically, the shape of an ellipse is denoted by
its eccentricity e, and to avoid getting embroiled in the mathematics
of conic sections (circles, parabolas and hyperbolas) all that is impor-
tant to know is that the larger the value of e, the more squashed is the
ellipse. The eccentricity of a circle is zero; for something the shape of
an egg it is about 0.5. Most of the planets, from Venus to Neptune, have
C
H
A
P
T
E
R

6
93
eccentricities of less than 0.06,

and so their orbits are indeed almost

circular. However, with an eccentricity of 0.254, Plutos orbit is quite
elongated and its heliocentric distance ranges between 29.6 and 48.9
AU. Neptune has an almost circular orbit with a mean heliocentric
distance of a little over 30 AU. This means that for about 20 years per
orbit, Pluto is actually closer to the Sun than Neptune. This situation
is unique; Pluto is the only planet to cross inside the orbit of another.
The existence of Pluto in a Neptune-crossing orbit is surprising. At
rst sight it would seem highly likely that at some time in the last few
billion years Pluto must have been crossing the orbit of Neptune at
just about the time Neptune was close to the crossover point. Neptune
is much more massive than Pluto, so the effect of such an encounter
on the larger planet would have been negligible, but the resulting
gravitational perturbation on Pluto would have been drastic. The
inevitable encounter should have moved Pluto into a new orbit,
perhaps sending it towards the Sun or ejecting it into deep space. The
reason why this has not occurred, and why Pluto has survived so long,
is that Plutos 245.96-year orbital period is almost exactly one and a
half times that of Neptunes 164.7-year journey around the Sun. This
situation, in which two objects have orbital periods that are simple
multiples of each other, is called a commensurability. In this case, the
relationship means that every time Neptune completes three orbits,
Pluto makes two. Consequently, at the end of this little celestial dance
the two planets come back into the same relative positions. Five
hundred or so years later, Pluto has made two more orbits and
Neptune another three, so once again the planets are back where they
started. If this special dynamical situation survives despite evolution
of the orbits with time, it is called a mean motion resonance. In this
case we say that Pluto is in Neptunes 2:3 mean motion resonance. So,
provided Pluto is in an orbit that happens to avoid close encounters
with Neptune, it will continue to avoid them on a regular basis. As it
happens, when Pluto is close to perihelion, and so is crossing
Neptunes orbit, Neptune is 90 degrees away around its orbit. This
huge distance is easily large enough to prevent a signicant perturba-
tion of Plutos orbit by Neptunes gravity.
Now such an exact commensurability is a very special situation. It
seems unlikely that one should occur, and even less likely that it could
survive over the age of the solar system. Over millions of years the
B
e
y
o
n
d

P
l
u
t
o
94

The innermost planet Mercury has an orbit with eccentricity of 0.2.

weak gravitational inuences of the other planets would be expected
to add up and disturb this special relationship. Indeed, the
PlutoNeptune commensurability could not exist in the real solar
system for very long but for the fact that there are other factors at
work. If Pluto is not in an exact commensurability, then over time its
orbit will change so that the crossover point drifts closer to Neptune.
When this happens Pluto begins to feel an increasing gravitational tug
from the much larger planet. The effect of this tug is to change Plutos
orbit by a small amount and to alter the orbital relationship between
it and Neptune. The effects of these little perturbations will eventu-
ally send Plutos orbit back the way it came. When this happens
Plutos orbit will begin to move back towards a condition of exact 2:3
commensurability. However, Pluto never manages to stay in this con-
dition since the orbit overshoots the stable point and begins to drift
towards that of Neptune from the opposite direction. This means that
Plutos orbit is now drifting towards Neptune from the other side and
the gravitational inuence of Neptune is now in the opposite sense.
Eventually, these perturbations will reverse the drift again. This will
send Plutos orbit back towards the exact resonance which it will
reach and then overshoot. Pluto will then begin to repeat its complex
orbital wanderings. Pluto never manages to remain quite in the 2:3
commensurability; instead, its orbit wanders back and forth, or
librates, around it with a period of about 70 000 years (a little less than
300 Pluto revolutions). So, while the ne details of Plutos orbit change
slightly over millennia, with small variations in inclination, eccen-
tricity and semimajor axis, these parameters oscillate gently around
some typical values. Fortunately for Pluto, they never undergo any
dramatic changes severe enough to jolt the planet into an unstable
orbit from which it can never return to the mean motion resonance.
The rst two trans-Neptunian objects found were on fairly circular
orbits well beyond the immediate gravitational inuence of Neptune,
but this was not true for the next four objects discovered. The prelimi-
nary orbits for 1993 RO, RP, SB and SC all seemed to indicate that they
were substantially closer to the Sun than 1992 QB
1
and 1993 FW. All
four seemed to be at heliocentric distances of around 3035 AU,
putting them right in the danger zone for potentially disastrous gravi-
tational encounters with Neptune. Since it was unlikely that many
objects could survive in this region for long, it was not long before the
idea that they might be in stabilising resonances surfaced. An early
hint of this came with the discovery announcement of 1993 RO, with
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
95
Brian Marsden noting that the object was located about 60 degrees
away from Neptune.
The signicance of this rather cryptic remark is that there are a
number of potentially stable points connected with giant planets.
The best known examples of these are the regions occupied by the so-
called Trojan asteroids, the rst of which was discovered in 1906. It
was named 588 Achilles, a character fromthe Iliad, Homers epic
poemabout the Trojan Wars. Achilles was found to have a mean dis-
tance fromthe Sun of about 5 AU, placing it at a very similar heliocen-
tric distance as Jupiter. This seems to be a very unlikely place to nd a
four and a half billion year old asteroid, but Achilles was able to
survive because the gravitational inuences of the Sun and Jupiter
combine to make the specic region of space it occupies fairly stable.
The existence of ve gravitationally stable points in a systemcontain-
ing two massive bodies and a third one of negligible mass was pointed
out by the eighteenth century mathematician Joseph Louis Lagrange,
and the stable points are named Lagrangian points in his honour.
Three of the Lagrangian points lie on a line connecting the two
bodies. The two others fall 60 degrees ahead of and behind the smaller
of the two large masses. As other asteroids were discovered in orbits
similar to that of Achilles it became traditional to name themafter
characters fromthe Iliad. The leading group contains characters
from(mostly) the attacking Greek army and the trailing group is pop-
ulated by the defenders of Troy. Because of the names given to the
objects that inhabit them, these stable regions are often referred to as
Trojan points. However, since the real solar systemcontains more
than two planets and one asteroid, the Lagrangian points are not
points at all, but regions. They are 1:1 mean motion resonances of
Jupiter.
Trojan asteroids are not xed in position as rigidly as some astron-
omy books suggest. They librate around their Trojan point and can
wander a fair way from it before drifting back. However, on average,
they are found in a region close to the magical 60 degrees from the
planet. The implication of Marsdens remark was clear, was 1993 RO a
Trojan asteroid of Neptune? The problem was that with such a small
observed arc, the orbit of 1993 RO, like those of the other new discov-
eries, was quite uncertain. What was needed was more observations,
preferably a few months hence, when the new objects had moved
somewhat and the observations could provide a bigger arc. However,
there was a Catch 22. To improve the details of their orbits, the objects
B
e
y
o
n
d

P
l
u
t
o
96
had to be observed when they had moved a fair way across the sky, but
without an orbit to predict their positions, could they ever be found
again? Luckily there was a clue. Dave Tholen had observed 1993 SC in
November 1993, and while his positions still allowed the object to be in
a variety of orbits ranging from 34 to 44 AU, Brian Marsden suspected
that, like Pluto, 1993 SC was in resonance with Neptune.
During the winter of 19934, the movement of the Earth meant that
the new objects slipped behind the Sun and for several months they
only rose above the horizon during daylight. This made recovery
observations of them impossible. It was not until early summer of
1994 that attempts to relocate them could begin. In May, Brian
Marsden published a set of predicted positions for the four objects
based on his assumption that their orbital stability was assured by
them being in 2:3 resonance with Neptune. These positions relied
heavily on this assumption and could have been in considerable error,
especially if the objects were in fact Trojan asteroids of Neptune.
The searching season opened in about June as the objects became
visible in the morning sky just ahead of the dawn. For a while there
was no news of them. This ominous silence was broken in the late
summer with the recovery of 1993 SC and 1993 RO from two observato-
ries. In August, Mark Kidger found the two objects using the Isaac
Newton Telescope in La Palma and Dave Jewitt and Jane Luu picked
them up in September from the 10 m Keck telescope on Mauna Kea.
Later in September, the Hawaii team recovered 1993 SB from the UH
2.24 m telescope. Although not sufcient to tie down the orbits of these
objects with very great precision, the new observations were all con-
sistent with orbits in the 2:3 Neptune resonance.

This fuelled specula-

tion that there might be a much larger population of similar objects in
Pluto-like orbits. According to Marsden, it was the realisation that
there were other objects in resonance with Neptune that nally
allowed Pluto to make sense; this was the key to understanding the
whole trans-Neptunian region.
The fourth object, 1993 RP, was never seen again, but the existence
of a population of objects in 2:3 resonance with Neptune was soon
conrmed. Note that the orbits of these objects are not all the same,
nor are they exactly the same as Pluto. The orbits have a range of
eccentricities and semimajor axes and the majority of them do not
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
97

None of the four turned out to be Neptune Trojans even though such locations should be
stable.
actually cross the orbit of Neptune (1993 SB is one of a few known
objects which do). The point is that they all come close enough to the
orbit of Neptune to have unstable orbits but for the effects of reso-
nances, which prevent a close encounter with the planet itself. The
existence of this population of Plutinos begged the question of
whether or not other of Neptunes mean motion resonances were also
populated. Theory indicated that there were other resonances; in par-
ticular there should be a population of objects with periods of close to
three hundred years in the 1:2 resonance. This resonance lies about
47.5 AU from the Sun, so objects in it will be fainter and harder to nd,
and it was not until 1996 that the rst such object (1966 TR
66
) was actu-
ally discovered.
However, the discovery of numerous objects in mean motion reso-
nances did not mean that all the trans-Neptunian objects were so sta-
bilised. In particular, the orbits of the rst two objects found, 1992 QB
1
and 1993 FW, were rather more distant than the Plutinos. Even
without the protection of resonances, they never approach Neptune
closely enough to be strongly perturbed by its gravity. These two
objects, and others like them, most closely resembled the population
of objects described by Edgeworth and Kuiper and so this region of
space became known as the classical Kuiper Belt. However, there
remained the dirty little secret of the Kuiper Belt; if this was the
source region of the short-period comets, what actually caused objects
to escape the belt and move into the inner solar system?
This problem was not one that could be solved directly by observa-
tion. In astronomy things usually happen very slowly and there is no
chance that any of the presently known trans-Neptunian objects will
evolve into comets during the lifetime of anyone who is observing
them today. Astrophysicists have similar difculties when they deal
with the evolution of stars and galaxies; here too, things usually
happen too slowly to follow during a single lifetime. Luckily for the
astrophysicists there are other ways of tackling their problems. Star
formation is going on throughout the galaxy and different star-
forming regions are of different ages. By observing different regions
astronomers can study snap-shots, if not moving pictures, of the star-
forming process. Astronomers studying extragalactic objects observe
objects so far away that the nite speed of light means that when they
observe distant galaxies they are peering backwards in time as well as
across vast reaches of space. This means that cosmologists can study
galaxies with a range of ages to try and understand how they form
B
e
y
o
n
d

P
l
u
t
o
98
and evolve. Solar system astronomers do not have either of these luxu-
ries. They are studying just one solar system (the recently discovered
extrasolar planets seem to be in systems rather different to ours) and
the distances are tiny by comparison with usual astronomical stan-
dards. The look-back time to the edge of the solar system is measured
in hours, not billions of years. If astronomers want to peer into the
distant past of our solar system, or project events far into its future,
their only recourse is to model it mathematically.
Since the effects of gravity are well understood, there is nothing
conceptually difcult about tracing the motion of one object under
the inuence of another. In principle it should also be easy to add in
further objects and see how mutual gravitational interactions affect
the overall picture. The problem is the sheer number of calculations
which need to be performed. As soon as one body moves, its effect on
all the others changes. Analytical techniques have been developed
over the centuries to get around this difculty, but these are usually
restricted to just a few objects. For example, analytical methods can
investigate the evolution of a couple of planets moving around a star,
but they cannot deal with a huge ensemble of particles all moving
about at the same time. It is only with the advent of computers able to
make many calculations in microseconds that detailed mathematical
modelling of the evolution of the solar system has been possible.
Modern computers allow theoretical astronomers to simulate the
motion of objects as they move through space under the inuence of
the combined gravitational forces of the planets and to determine the
effects of the competing gravitational tugs on the evolution of these
orbits. This process is known as integrating the orbital elements to
see what they will look like some time in the future, or what they were
some time in the past. Integrating orbital elements for real objects
needs to be done so their positions can be calculated accurately
enough for observations with large telescopes (it is especially impor-
tant if the object happens to have passed near another planet recently
and may have had its orbit changed signicantly during the
encounter) but orbital integrations can be used as a research tool in
their own right.
One approach is to take the details of an object that actually exists
and to make some small, but signicant, change in one aspect of its
orbit before allowing the computer to track how it might then evolve.
For example, take the orbit just as it is, but move the object along its
path a little. This will make it seem as if one object had suddenly
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
99
jumped forward a few years while everything else stood still. The
effect of this can vary from not very signicant, which suggests the
orbit is fairly stable, to quite dramatic. For example, Iwan Williams of
Queen Mary and Westeld College in London, one of the group which
had discovered 1993 SB and 1993 SC, considered the evolution of the
orbit of 1993 SB over a ten million year period. He found that even
though it crossed the orbit of Neptune, and its orbit did oscillate grad-
ually with time, the orbit was stable during this period and never
changed signicantly. However, if he kept all the orbital details the
same and just assigned them to a date ve years hence (and remember
this is only 2% of 1993 SBs 250 year orbital period) the result was
rather different. The main effect was to change the geometry when
1993 SB made its closest approaches to Neptune. The effect of this was
to cause very signicant changes in all its orbital parameters. While
the 1993 SB clone survived the ten million year integration, it was
clear that something dramatic was going to happen to it, and probably
sooner, rather than later.
This result reveals an important point about the stability of these
resonant orbits; while objects in them might well be quite stable, even
slightly different orbits may be unstable. What we see today is not an
intact population of ancient objects, but just the lucky few survivors
that happened to nd themselves in stable orbits, or were somehow
pushed into them before they could be ejected from the solar system.
This is why Pluto has survived to this day. It is not just a fortunate
coincidence that its orbit is in a mean motion resonance with
Neptune; if it had been otherwise Pluto would not have survived long
enough for Clyde Tombaugh to discover it.
Another way of using computers to study solar system evolution is
the one taken by theoretical physicists like Martin Duncan and Hal
Levison, who we last encountered in chapter 4 unsuccessfully trying
to detect the Kuiper Belt from a telescope in Arizona. They, and others,
have developed very detailed computer codes to follow the evolution of
objects initially in orbits close to Neptune. They can then compare the
outcome of their simulations with what is actually observed out there
in real, as distinct from cyber, space. To do this, they set up models in
which hundreds, or sometimes thousands, of test particles (essen-
tially theoretical Kuiper Belt objects), with a wide range of orbital
parameters, are set in motion. Their orbits are then integrated over
very long periods. The details vary from simulation to simulation,
depending on the objective of each experiment, the cleverness of the
B
e
y
o
n
d

P
l
u
t
o
100
code written to do the calculations and the amount of computing
power available. For example, in experiments conducted around 1992,
the orbits of 1000 test particles were integrated at 1 or 10 year inter-
vals over a period of 1000 million years. This is equal to about one
quarter of the age of the solar system and should have given a reason-
able idea of how the nal result would come out if there had been time
to allow the simulation to run longer. Of course, some approximations
had to be made the test particles were assumed to be without any
mass, so mutual interactions could be ignored. Also, not all the
planets were included, as the inuence of the gravity from the small
and distant inner planets is tiny compared with that of the four giant
planets. In fact, when studying objects beyond Uranus, even the effects
of Jupiter and Saturn can often be ignored, provided their effects on
the orbits of Uranus and Neptune are taken into account before calcu-
lating the gravitational inuences of Uranus and Neptune on the test
particles.
No computers yet exist that can integrate objects in every conceiv-
able orbit over the age of the solar system with high enough temporal
and spatial resolution to produce any kind of exact answer. Even if
there were such machines, it would still not be possible to complete
the job. The effect of non-linear dynamics, or chaos theory, makes it
impossible to predict things in the real world with complete accuracy.
To see why this is so, consider that in a real solar system one of the
particles might be hit by a meteorite which could change its orbit very
slightly. This minor incident might not make any difference to the
outcome, but it might cause a slightly different effect on the path of
another particle. This in turn might have knock-on effects elsewhere.
Before long, astronomically speaking, these effects could ripple
through the entire population and produce an outcome quite different
from the one which would have resulted if the rst particle had not
had that meteorite impact. Such subtle effects cannot be fully mod-
elled, so innite precision is never going to be possible in these
dynamical simulations.
Since it is impossible to do everything, detailed studies of the effect
of specic parameters on the outcome of a simulation are done by
starting with test particles having a restricted range of some parame-
ters and a complete mixture of all the others. For instance, to investi-
gate the effect of semimajor axis on the outcome of a test, a selection
of orbits all having fairly low inclinations and a range of eccentrici-
ties might be chosen. To see what happens to particles at high
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
101
inclinations, the initial set might include only a small range of eccen-
tricities, but a wide range of inclinations. Once these initial values
have been chosen, then the particles can be put into these orbits at
random positions, the orbits distributed around the Sun and the simu-
lations started. In most simulations, once a test particle has reached
some critical limit, for example when it comes close enough to
Neptunes orbit to have a dramatic encounter, or it has moved more or
less out of the gravitational reach of the planets, it is removed from
the simulation to save computing power.
This is a very different kind of astronomy from that practiced by
Jewitt, Luu, Offutt and the other observers of the real Kuiper Belt. It
does not involve travel to exotic destinations such as Hawaii, Chile or
the Canary Islands and it does not even necessarily involve staying up
at night. It is a job where the tools are computer workstations not tele-
scopes and where capability is measured in terms of processor power,
not mirror size, and Megaops (millions of oating point operations
per second) rather than elds of view and limiting magnitude.
Martin Duncan has been very active in this esoteric eld and he is
now a professor of physics at Queens University in Kingston, Ontario.
Like many university academics, he ts his research in around teach-
ing and general administration of his department. His day often
starts early, with a check on whatever computer runs are going on at
that moment. His group tends to set up long simulations, which may
take weeks or months to nish, but he says that it is nice to be reas-
sured that everything is running smoothly. Some of his simulations
involve dozens of machines and many months of computing time to
build up the statistics of how the virtual solar systems are forming
and evolving. While, as he puts it, his workstations are just sitting
there doing lots of simple calculations over and over again and proba-
bly getting very bored in the process, Duncan has time to think about
what it all might mean. He has to consider what further tests need to
be done and to try and have his next good idea ahead of the competi-
tion. Like the observers who have to compete for telescope time, the
theoretical physicists are also competitive people. While they recog-
nise and respect the good work of others, they also sometimes knock
themselves on the head and ask themselves, Why didnt I think of
that?.
What the computer simulations of Martin Duncan and other
people doing similar projects have shown is that the structure of the
region around Neptune is very much more complicated than origi-
B
e
y
o
n
d

P
l
u
t
o
102
nally thought. Objects in near-circular orbits can appear to be per-
fectly stable for hundreds of millions of years, with just gentle, cycli-
cal changes in their orbital elements over ten million year timescales,
then suddenly the gradual build-up of the gravitational inuences of
the planets conspire to produce a rapid change. Over a few million
years, the once apparently stable orbits can change rapidly until they
become Neptune crossing, or at least get close enough to Neptune to
suffer a major gravitational perturbation, and be ung out of the
trans-Neptunian region completely.
By studying the fate of thousands of ghostly objects, Hal Levison
and Martin Duncan have been able to map out the likely structure of
the trans-Neptunian region. It is a strange kind of map. It does not
plot positions in space, or motion around the Sun, but islands of sta-
bility in the strange mathematical space of orbital elements. An
example of one of these diagrams is shown in gure 6.1. This shows
that the trans-Neptunian region is expected to have a complex struc-
ture, although one whose general features are fairly easy to under-
stand. In general, any object in an orbit that brings it to within about
35 AU of the Sun will be removed by the cumulative effects of
Neptunes gravity quite quickly. The models show that this region is
expected to be fairly empty. Any objects found in such orbits today
must have drifted into them quite recently and are already doomed.
For them, a dramatic gravitational encounter with Neptune and per-
manent removal from the Kuiper Belt is only a matter a time. The only
exceptions to this rule are objects such as the Plutinos whose orbits
are librating around Neptunes mean motion resonances. They appear
to be stable for the age of the solar system.
Moving outward, there is a region from about 40 to 42 AU in which
orbits are highly unstable. This is a region where another kind of res-
onance, a so-called secular resonance, becomes important. We saw
that for a mean motion resonance to occur the orbital periods of two
objects need to be a simple ratio of each other, like the 2:3 relationship
between Neptune and Pluto. If the orbits of the planets remained
xed in space, then that would be all there was to worry about.
However, in reality, the mutual gravitational interactions of the
planets cause their orbits to precess. Precession is a process in which
the direction dened by the long axis of an elliptical orbit drifts
slowly around the Sun. It is similar to the way in which the hoop of a
hula hoop dancer drifts gradually around the dancers waist. Provided
the rates of precession of any two orbits are not the same, then an
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
103
F
i
g
u
r
e
6
.
1
T
h
i
s

g
u
r
e

s
h
o
w
s
t
h
e

p
r
e
d
i
c
t
e
d

d
y
n
a
m
i
c
a
l

l
i
f
e
-
t
i
m
e
s

o
f
t
r
a
n
s
-
N
e
p
t
u
n
i
a
n
o
b
j
e
c
t
s

w
i
t
h

v
a
r
i
o
u
s

i
n
i
t
i
a
l
o
r
b
i
t
a
l

c
h
a
r
a
c
t
e
r
i
s
t
i
c
s
.
T
h
e
o
r
i
g
i
n
a
l

o
r
b
i
t
s

o
f
t
h
e

t
e
s
t

p
a
r
t
i
-
c
l
e
s

a
r
e

d
e

n
e
d

b
y

s
o
m
e

c
o
m
b
i
-
n
a
t
i
o
n

o
f
i
n
i
t
i
a
l

e
c
c
e
n
t
r
i
c
i
t
y
a
n
d

s
e
m
i
-
m
a
j
o
r

a
x
i
s

a
n
d

t
h
e
s
i
m
u
l
a
t
i
o
n

i
s

a
l
l
o
w
e
d

t
o
p
r
o
c
e
e
d
.
T
h
e

c
u
r
v
e
d

l
i
n
e
s
m
a
r
k
e
d

b
y

q
=
3
0

a
n
d

q
=
3
5

s
h
o
w
t
h
e

c
o
m
b
i
n
a
t
i
o
n
s

o
f
e
c
c
e
n
t
r
i
c
-
i
t
y

a
n
d

s
e
m
i
-
m
a
j
o
r

a
x
i
s

w
h
i
c
h
a
l
l
o
w

o
b
j
e
c
t
s

t
o

c
o
m
e

w
i
t
h
i
n

3
0
a
n
d

3
5

A
U

o
f
t
h
e

S
u
n

a
t

t
h
e
i
r
c
l
o
s
e
s
t

a
p
p
r
o
a
c
h

(
p
e
r
i
h
e
l
i
o
n
)
.
N
o
t
e

t
h
a
t

a
l
m
o
s
t

n
o

o
r
b
i
t
w
i
t
h
i
n

(
t
o

t
h
e

l
e
f
t

o
f
)

t
h
e

q
=
3
0
A
U

l
i
n
e

i
s

s
t
a
b
l
e

o
v
e
r

t
h
e

a
g
e

o
f
t
h
e

s
o
l
a
r

s
y
s
t
e
m
.
V
a
r
i
o
u
s
N
e
p
t
u
n
e

m
e
a
n

m
o
t
i
o
n

r
e
s
o
-
n
a
n
c
e
s

a
r
e

m
a
r
k
e
d

b
y

v
e
r
t
i
c
a
l
l
i
n
e
s
,
s
u
c
h

a
s

t
h
e

1
:
2

r
e
s
o
n
a
n
c
e
a
t

a
b
o
u
t

4
7
.
8

A
U
.
T
h
e

s
t
a
b
l
e
(
l
i
g
h
t

c
o
l
o
u
r
e
d
)

r
e
g
i
o
n
s

t
e
n
d

t
o
b
e

c
l
o
s
e

t
o

t
h
e
s
e

r
e
s
o
n
a
n
c
e
s
.
object outside a precessing planet experiences a gravitational force
which is symmetrical. Over a suitably long timescale the changing
gravitational inuences of the interior planet average out. If,
however, the period of precession of one object is the same as that of
another, then the situation is not symmetrical and there can be
signicant gravitational torques on the smaller of the two objects.
The effect of these forces can be to change the eccentricity of the
smaller objects orbit. In a solar system like ours, in which all the
bodies are not orbiting in exactly the same plane, additional secular
resonances can occur. Each planets orbit crosses the invariable plane
of the solar system at two points, called the ascending and descending
nodes. The position of these nodes also precess and if the rate of pre-
cession of the nodes of two orbits are the same then gravitational
forces can change the inclination of the smaller bodys orbit. These
types of resonances are normally denoted by the greek letter or
1
respectively, followed by the number of the planet doing the perturb-
ing. The secular resonances with Neptune, which has the dominant
effect on the trans-Neptunian region because of its proximity, are
denoted
8
and
18
.
It so happens that the region around 4042 AU is where several of
the secular resonances of Neptune and Uranus occur, and these con-
spire to make this region unstable. Any objects which have the misfor-
tune to wander into this part of the solar system are rapidly removed.
The secular resonances also destabilise objects in high inclination
orbits with semimajor axes close to that of Pluto, explaining why the
stable Plutinos are not far from the plane of the solar system. Beyond
about 42 AU these secular resonances cease to be important and a
fairly stable region of orbital space exists. Objects here can often
survive for the age of the solar system without needing to protect
themselves by remaining in mean motion resonances of Neptune.
Using their computer models the theoretical astronomers, or
orbital mechanics

as they are sometimes jokingly referred to by

observational astronomers, were able to paint a self-consistent picture
of the space beyond Neptune. Their calculations showed that it is a
region which has been dynamically sculpted by aeons of gravitational
interactions into a complex structure. Some regions were expected to
contain many objects while others should be almost empty. However,
this was all theory; the proof of the pudding was to be in the eating.
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
105

Strictly, they are orbital mechanicians.

Would the theory survive a head-on crash with real observations? To
nd out, astronomers plotted the real trans-Neptunian objects onto
the map of stable regions to see how well the actual distribution
matches the predicted one.
The answer, which is shown in gure 6.2, is that it ts quite well.
Around about 39 AU there is a population of Plutinos with low-
inclination orbits stabilised by the 2:3 mean motion resonance. The
4042 AU gap, which theory predicts should be cleared by secular reso-
nances, is indeed empty. Beyond 42 AU are a number of objects spread
more or less evenly across the predicted stable region. Only a few
objects are seen beyond 45 AU, but this may be a selection effect. At
these large distances objects will be that much fainter and corre-
spondingly harder to detect. There is, however, one anomaly. Theory
predicts that there is a stable region for low-inclination orbits with
semimajor axes in the range 3639 AU. This is comparatively close,
closer than the Plutino zone, and so it should be quite easy to nd
objects there, but in fact none have been discovered so far. How can
this be explained?
B
e
y
o
n
d

P
l
u
t
o
106
0
0.1
0.2
0.3
0.4
30 35 40 45 50
N < 1 Opposition
N 1 Opposition
O
r
b
i
t
a
l

E
c
c
e
n
t
r
i
c
i
t
y
Semi-Major Axis [AU]
q = 35 AU
2
:
1
3
:
2
4
:
3
5
:
3
X
q = 30 AU
Figure 6.2 The orbital distribution of real trans-Neptunian objects. Note the concentra-
tion of objects in the 2:3 mean motion resonance. These are the Plutinos. Pluto itself is
marked by an X just above the line representing objects with a perihelion distance (q) of
30 AU. (Dave Jewitt.)
There are at least two possibilities. Orbits in this region are only
stable if they are fairly circular, i.e. they have eccentricities of less
than about 0.05, and have low inclinations. If some process caused the
orbits in this regions to be systematically altered, to have either larger
eccentricities or higher inclinations, then the objects would have been
ejected by now and the region would indeed be empty. So perhaps the
eccentricities and/or inclinations of objects in this region were
pumped up to higher values by mutual interactions between the
objects themselves. The presently known trans-Neptunian objects are
not massive enough to do this, but this inner region may once have
contained a few larger objects. This is not by any means impossible;
Pluto, its moon Charon and Neptunes largest satellite Triton, are
bigger than typical trans-Neptunian objects and all are found in this
vicinity. Perhaps a few large objects once existed here and were
responsible for gravitationally clearing the region before being
removed by collisions or some other as yet unexplained process.
If the idea of creating a few large objects which can clear out the
3639 AU region and then themselves vanish does not appeal, there is
another less ad hoc explanation. This hypothesis might bear not just
on the clearing of the 3639 AU region, but on the trapping of Pluto
and the other objects in the 3:2 resonance. Perhaps the structure of the
solar system we see today is not the same as it was early in its history.
In particular, it is possible that changes in the orbit of Neptune might
have played a role in sweeping out the 3639 AU region and trapping
objects in mean motion resonances. This sort of effect would not be
found by simulations which assume that the orbits of the giant
planets are the same now as they were over four billion years ago. If
the planets orbits once moved around the solar system, then the
effects of this evolution will be seen in the present structure of the
trans-Neptunian region. Studying the Kuiper Belt today can provide
important clues to how the outer solar system formed.
Can planets orbits migrate across the solar system? Yes they can
and what is more they probably did. During the nal stages of the for-
mation of the solar system a situation must have existed when the
four largest planets had more-or-less nished growing and were
already on fairly circular, low-inclination orbits. Any nebular gas
which had not been incorporated into planetesimals had long since
vanished, but large numbers of planetesimals were still to be found in
the space between the planets. They were particularly common in the
region beyond Neptune where no massive planet had formed. One by
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
107
one, these planetesimals would have encountered Neptune and, while
a small fraction would have crashed into the planet, the majority
would have been scattered into new orbits by Neptunes gravity.
However, gravitational scattering is not a process that only affects the
planetesimal. The principle of conservation of angular momentum
means that the orbit of the scattering planet must also be affected. For
a single interaction, the effect on the orbit of the giant planet is negli-
gible, but if many encounters are occurring the effects can build up
over time. The combined effects of many such interactions may
signicantly change the orbital energy and angular momentum of the
scattering planet.
Early during the scattering process, when there were many plan-
etesimals, there would have been an equal number of objects scat-
tered inwards and outwards. The effect of these events on the
scattering planets orbit would cancel out, leaving no net change in its
energy. Later on, this situation would change. The ultimate fate of the
scattered planetesimals depended on which direction they went after
their rst encounter with Neptune. Most of the objects scattered out-
wards from Neptune could return to encounter the planet again in the
future, but this is not true of objects sent inwards. As shown by Ed
Everhart, objects scattered inwards are likely to fall under the gravi-
tational control of the other giant planets and eventually reach the
sphere of inuence of massive Jupiter. If this happens there is a good
chance that Jupiters gravity will eject them from the solar system
entirely. So, late in the scattering process, most of the objects interact-
ing with Neptune are the survivors of objects thrown outwards on
their rst encounter and returning Sunwards. There are compara-
tively few objects interior to Neptune diffusing outwards. Towards the
end of the process the energies of the objects reaching Neptune are
biased, the average energy of each gravitational interaction is no
longer zero. The encounters are now, in effect, a drag on Neptune and
this acts to increase the orbital radius and causes the planet to drift
outwards. Uranus and Saturn, which are also outside Jupiter, would
have suffered similar, although smaller effects. They would also have
migrated outwards slightly. Jupiter, the most massive planet, having
indirectly provided the energy to move the other planets outwards by
ejecting many planetesimals entirely, would have moved inwards to
conserve the solar systems total angular momentum.
These changes in the diameter of Neptunes orbit, which might
have amounted to several astronomical units, would have a profound
B
e
y
o
n
d

P
l
u
t
o
108
effect on the outer solar system. As Neptune moved outwards, its
orbital period would have changed. As it did so, the positions of its
mean motion resonances would also have moved. The resonances
would have drifted outwards ahead of the planet and moved through
the region occupied by the surviving outer solar system planetesi-
mals. The idea of slowly moving resonances sweeping up planetesi-
mals as they went along has been developed by another of the small
band of theoretical astrophysicists trying to sort out the 4.5 billion
year history of the solar system.
Renu Malhotra was born in New Delhi, India, in 1961. She grew up
mostly in Hyderabad, a city more or less in the geometric centre of the
Indian sub-continent. Looking back, she says that it seems she was
fated to study physics. As a child she had many conversations about
nature, the physical sciences and the atmosphere with her father, an
aircraft engineer with Indian Airlines. These led her to read physics
at the Indian Institute of Technology in Delhi. This is one of ve
major engineering campuses in India and one to which entrance is
very competitive, with only about 1000 of the 200 000 potential stu-
dents being admitted each year. After passing the nationwide
entrance examination, and studying there for ve years, Malhotra
had gained what she describes as an excellent education.
From Delhi, she went to the United States. She entered graduate
school at Cornell University at Ithaca, in New York State, to study for a
PhD in physics. It was here, while working in the eld of theoretical
non-linear dynamics, that she discovered solar system dynamics and
soon became interested in the problems it raised. After all, she says,
The solar system is the pre-eminent dynamical system, all of modern
physics came out of the study of planetary motions. To Renu
Malhotra the study of the dynamics of solar system seemed like a
natural home for the things she loves to do, combining mathematical
analysis and theoretical physics to understand the world around her.
She gained her doctorate in 1988 then went to the California Institute
of Technology in Pasadena for a period of postdoctoral research. She
then moved to the Lunar and Planetary Institute in Houston, Texas.
Like Martin Duncan and Hal Levison, she spends her working time
reading scientic papers, doing analytical studies of dynamical prob-
lems and running computer models on her desktop workstation.
Inside the machine millions of years ow past in hours, as the primi-
tive solar system evolves. Electronic planets grow and their orbits
change as phantom collisions occur by the thousands. Eventually,
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
109
some order emerges and a model of how the the solar system might
appear is revealed, ready to be compared to real planets, moons and
asteroids in our planetary system. Malhotra rst used the idea of res-
onance sweeping to explain how Pluto might have been captured into
its present orbit. She then expanded this idea to account for the trap-
ping of other objects into planetary resonances. Her results suggest
that this sweeping would be quite sufcient to clear out planetesimals
from the region inside about 39 AU. As the orbit of Neptune expanded,
objects would fall into the resonances and would then be moved
slowly outwards along with the planet. A consequence of this process,
which can be tested by observations, is that not only must the regions
swept clear be empty, but the resonances must contain more objects
than could have got there by chance if the resonances had not
migrated through the outer solar system, picking things up as they
went along.
Although the resonance-sweeping hypothesis does explain the lack
of objects inside 39 AU, and the large population of objects in the 3:2
resonance, it is not without its problems. For one thing, it predicts
that the 2:1 Neptune resonance should also be heavily populated. If
Neptune did migrate outwards, then its 2:1 mean motion resonance
would have started out in a region containing large numbers of
objects. These objects ought to have been picked up as the resonance
moved outward. Indeed, if resonance sweeping was effective, then not
only should the 2:1 resonance be quite full, but the area across which it
swept should be quite empty. Unfortunately, this does not appear to be
the case. Not only is the original location of the 2:1 resonance still
occupied by large numbers of classical Kuiper Belt objects, but the
present 2:1 resonance seems to be fairly empty. This may be because
the objects there have escaped somehow, or because of some observa-
tional bias in the way objects have been discovered so far, or perhaps
because the resonance sweeping did not occur exactly as the models
postulate. In particular, the timescale of the resonance sweeping may
be quite critical. The Japanese theorist Shigeru Ida has shown that
the 2:1 Neptune resonance can be unpopulated if the orbital migration
of Neptune happens in less than 5 million years or so.
Despite the difculties with each model, and such problems are to
be expected in a subject which is still in an early stage of development,
it seems that a reasonable picture of how both the Plutinos and the
classical Kuiper Belt beyond 42 AU can be sketched out. However,
there remains another type of trans-Neptunian object to explain.
B
e
y
o
n
d

P
l
u
t
o
110
From where did 1996 TL
66
, the object discovered by Jane Luu and
others then assiduously recovered by Carl Hergenrother and Warren
Offutt, come?
The existence of a population of objects on highly eccentric orbits
which only graze the planetary system when they are at their closest
to the Sun was rst put forward by Julio Fernandez and Wing Ip when
Fernandez was working in Germany during the early 1980s. More
recent work, particularly numerical integrations of objects encoun-
tering Neptune done by Martin Duncan and Hal Levison in the late
1990s, showed that while most objects encountering Neptune had
fairly short lifetimes, about 1% of the objects that they started with
were still in orbits beyond Neptune after 4 billion years. It turned out
that these objects could survive for such a long time because during
their wanderings they were likely to spend quite a bit of time tem-
porarily trapped in distant mean motion resonances with Neptune.
This process can be illustrated through the medium of the numerical
integrations by following one of these imaginary particles around the
solar system. For 70 million years the object wandered randomly
about near Neptune before it was trapped in Neptunes 3:13 mean
motion resonance. It stayed here for 50 million years and then
escaped, spending a little under 200 million years in limbo until it fell
into the 4:7 mean motion resonance. It remained there for a further 340
million years. Leaving this resonance it then drifted into the 3:5 mean
motion resonance where it stayed for about 500 million years. The test
particle then escaped that resonance and wandered about the outer
regions of the solar system for the remainder of the simulation.
If objects can survive this long, and if there were a signicant
number of objects in the UranusNeptune region soon after the solar
system formed, it seems that there could be a large number of objects
in this scattered disc today. This conclusion means that the rst scat-
tered disc object, 1996 TL
66
, was not unique, it was just the harbinger
of a much larger population of objects waiting to be discovered. More
scattered disc objects have since been found but it may be a little while
before this population can be mapped in detail. With orbits extending
to 200 AU, most of the objects in the scattered disc will spend most of
their time a great distance from the Sun. For much of the time they
will be very faint. With present technology, only a few objects have yet
been found when they were beyond 50 AU and an object at a 100 AU
will be sixteen times fainter still. For the moment at least, only scat-
tered disc objects that happen to be close to perihelion will be found by
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
111
current search programmes. The outer regions of the disc are likely
to remain unknown territory for a while.
Before we leave this arcane world of phantom particles, chaos and
resonances, it is worth taking a moment to review the state of our
understanding this complex region. How does what has been learned
relate to the problem of the origin of the short-period comets?
Numerical integrations and analytical studies are in general agree-
ment that the inner edge of the trans-Neptunian region occurs at
about 35 AU from the Sun. Any object which nds itself inside this
region cannot survive there for long before its orbit is drastically
modied by the gravitational inuences of the giant planets. The
region from 35 to 42 AU is complex and the population here has been
subjected to considerable dynamical evolution. Regions of stability
exist around the mean-motion resonances. Objects librating around
these locations avoid close approaches to Neptune and so can survive
for the age of the solar system. Outside the mean-motion resonances,
most orbits are not stable and these regions have long since been
cleared of any primordial planetesimals. At the boundaries of the res-
onances are semi-stable states in which objects can survive for bil-
lions of years, but which are not protected forever. From time to time
B
e
y
o
n
d

P
l
u
t
o
112
Figure 6.3 The orbits of the best observed scattered disc objects. The small circle in the
centre is the orbit of Jupiter. The gure is a square with sides of 400 AU. (Chad Trujillo.)
objects in these regions stray just a bit too far from safety and escape
from the resonances. Once they do they may begin a journey towards
the inner solar system and be destined to become a Jupiter family
comet, or suffer a dramatic collision with one of the planets.
Beyond about 42 AU is a stable region where most orbits are fairly
circular and of low inclination. This region contains objects that have
probably been there for the age of the solar system and is the closest
thing yet found to the disc of primordial material proposed by
Edgeworth and Kuiper half a century ago. Ironically, it is fairly
unlikely that objects from here will ever enter the inner solar system.
Even though some of the orbits are chaotic, they tend not to evolve
into paths which can encounter Neptune. Mixed in with these bodies
is another population, the scattered disc. This comprises objects on
more highly inclined and more eccentric orbits that dip inside the
classical Kuiper Belt, but spend most of the rest of their time much
further away, travelling a few hundred astronomical units from the
Sun. These are objects which were originally formed in the
UranusNeptune region and were subsequently ejected to great dis-
tances. However, orbits in the scattered disc are not stable over
timescales of billions of years. Although they may spend signicant
amounts of time in outer mean motion resonances, scattered disc
objects also have periods in which their orbits undergo chaotic wan-
dering and which may bring them back to a close encounter with
Neptune and perhaps a one-way ticket to the inner solar system.
Beyond about 50100 AU remains uncharted territory, where the gravi-
tational effects of the planets are not signicant, and to which we will
return.
S
o
r
t
i
n
g

o
u
t

t
h
e

d
y
n
a
m
i
c
s
113
What are little
planets made
of ?
Wherever they come from, it has long been recognised that comets are
very old and that they offer a way of studying the chemical composi-
tion of the early solar nebula. The recent history of the solar system is
written in the atmospheres and surfaces of the planets and their
moons, but most of the chemical evidence has long since been
removed. More than four billion years of heating, asteroid impacts,
geology and weather have altered the planets almost beyond recogni-
tion. Comets, on the other hand, have been preserved in deep freeze
for billions of years and have probably experienced little thermal
processing since they were formed. They are natural time capsules
preserving a record of the long-vanished solar nebula. Many an
application for telescope time has begun with words to the effect of,
Comets represent the least processed solar system material available
for study, but most of the people writing these applications, and
perhaps even a few of those awarding them telescope time, must have
known that this was not completely true.
The problem is that before a comet can be studied, it has rst to be
discovered. Comets are usually found by eagle-eyed observers who
detect the faint hint of a diffuse coma around an otherwise starlike
object. So, before a single detailed observation of it can be made, the
comet has already started to change. The very thing which makes a
comet bright enough to discover is the sublimation of the ancient ices
which have survived on or near its surface since it was formed.
Observations of comet HaleBopp, and of the Centaur Chiron, show
that cometary activity is not restricted to objects close to the Sun.
Comet HaleBopp had a distinct coma when it was more that 7 AU
away and Chiron has been seen to outburst at even greater heliocen-
tric distances. What this means is that active comets are not the least
C
H
A
P
T
E
R

7
115
processed material in the solar system at all. That sobriquet surely
belongs to trans-Neptunian objects, which have never felt the warmth
of the Sun. It is this extreme age which makes observations of them so
important. Unfortunately, the trans-Neptunian objects are quite small
and very distant, a combination which makes them faint and difcult
to study with all but the largest telescopes.
The rst clues to the physical nature of the trans-Neptunian
objects came during the early searches by Jane Luu and Dave Jewitt.
The surveys were done using a lter which isolated mostly red light
B
e
y
o
n
d

P
l
u
t
o
Figure 7.1 The summit of Mauna Kea in Hawaii. Most of the physical studies of the
trans-Neptunian population have been done from these observatories. At the extreme left
is the tiny dome of a 0.6 m telescope, then the silver dome of the UK Infrared Telescope is
partly hidden behind the University of Hawaiis 2.24 m telescope. The multi-national
Gemini telescope has its dome slit open. Next to this is the 3.6 m CanadaFranceHawaii
telescope. At the rear of the summit ridge are the Japanese Subaru telescope and the twin
domes of the W.M. Keck observatory. The NASA Infrared Telescope Facility is the silver
dome projecting from the square building centre right. In the valley are two submillimetre
telescopes; the Caltech Submillimetre Observatory and the James Clerk Maxwell
Telescope (JCMT). The building to the right of the JCMT is the maintenance facility for
an array of small submillimetre telescopes which had not been installed when this photo-
graph was taken. (Richard Wainscoat.)
since the combination of this lter and the response of their CCD
gave the highest possible sensitivity. However, once some objects had
been found, they were soon observed through other lters to deter-
mine their colours. The objective was to see where the trans-
Neptunian objects tted in the general scheme of minor planets,
comets and so on. An obvious question was, Are the Kuiper Belt
objects very red like Pholus, more-or-less neutral like Chiron, or are
they some colour in between?. The rst observations were not conclu-
sive. 1992 QB
1
had a fairly red colour, but 1993 FW seemed to be
neutral.
This rst hint that the colours of the Kuiper Belt objects might be
diverse was soon conrmed. Luu and Jewitt reported that there was a
wide range of colours amongst a dozen or so trans-Neptunian objects
they had observed. A similar conclusion was reached by Kent
Universitys Simon Green and other British observers who had been
observing from La Palma. A European group led by Antonietta
Barucci of the Meudon Observatory in Paris agreed. This was rather
odd, since it suggested that the surface compositions of these suppos-
edly primitive objects were not the same. Perhaps the surfaces of the
trans-Neptunian objects were not ancient after all?
The conventional wisdom was that the Kuiper Belt would contain
objects which were predominantly icy. They were expected to com-
prise a mixture of frozen gases such as carbon monoxide, ammonia
and methane together with water ice and relatively small amounts of
dust. However, while pure ices are blueish in colour, it is unlikely any
ancient ices in the trans-Neptunian region would be pure. Their
extreme ages would presumably have exposed them to at least some
chemical modication. Although at such great distances there is not
much solar radiation, indeed the Sun would appear little more than a
bright star, Kuiper Belt objects would certainly have been bombarded
by a steady ux of high-energy cosmic rays for billions of years.

Cosmic rays have sufcient energy that they are capable of breaking
individual chemical bonds and when they encounter icy surfaces they
smash their way a few metres into the material and do just that, break-
ing the bonds which hold together the simple molecules they
encounter. The fragments of molecules which are produced then
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
117

Cosmic rays are not rays at all, but charged particles of various types. They are mostly
protons and electrons plus a few heavier ions which have been ejected from violent
events elsewhere in the galaxy and are travelling through interstellar space at great
speed.
recombine in unpredictable ways to form more complex compounds
of larger molecular weight. The large organic molecules formed by
this process tend to give the material a reddish hue, but the material
does not stay red forever. Hydrogen atoms liberated during the
process, being small, are able to escape from the solid material and do
not always remain long enough to get chemically bound up again. So,
if the bombardment continues, the icy surface becomes relatively rich
in carbon. This excess of carbon, being black, tends to make the mate-
rial dark. The mish-mash of chemical bonds formed also makes the
material refractory, or hard to evaporate. The result is that a tough
crust called an irradiation mantle is formed. No-one really knows how
long it takes to make a refractory crust that is stable for long periods,
but estimates tend to run in the range of 10100 million years. Such a
long timescale is impossible to duplicate in the laboratory, although it
can be simulated by using very large doses of radiation over shorter
periods, but it is quite short in terms of the age of the solar system.
This suggests that all the Kuiper Belt objects ought to have refractory
crusts and so be dark and uniform in colour. So, at rst sight, a wide
range of colours is rather hard to explain.
However, science is about trying to understand the unexpected and
it was not long before several ideas were put forward to explain the
apparent differences amongst these supposedly similar objects. The
simplest suggestion was that the chemical composition of the objects
was not the same to start with. This is by no means impossible; there
are distinct trends in composition with heliocentric distance amongst
other groups of solar system objects. For example, in the main aster-
oid belt between Mars and Jupiter, the innermost objects are more
rocky and have less water bound up in them than those found further
out. These trends are attributed to the relatively rapid fall in the tem-
perature of the solar nebula at increasing distances from the Sun. The
temperature gradient would have tended to drive volatile ices away
from the inner edge of the asteroid belt, leaving behind dusty material
which went on to form mostly rocky objects. However, this sort of
mechanism is less likely to have been signicant in the trans-
Neptunian region. At these great distances from the Sun the range of
temperature across the planetesimal formation region would have
been fairly small, only 10 degrees or so. None the less, if objects in the
trans-Neptunian region originally formed at a range of heliocentric
distances before being transported outwards, by resonance sweeping
or some other gravitational effect, then perhaps it is naive to believe
B
e
y
o
n
d

P
l
u
t
o
118
that they all have the same initial composition. Adding to the mystery
is that there is no agreement about trends in the colours with any
orbital parameters. The different coloured objects all appear to be
mixed together.
Another alternative explanation of the colour diversity is to invoke
some kind of resurfacing which gradually changes the colours of the
objects. This could happen if the surface was slowly developing a
reddish colour under the inuence of cosmic rays and solar radiation
when something forced out fresh material of a different colour from
inside. By analogy with the Centaur Chiron, which is fairly neutral in
colour, cometary activity might be the cause of this. However, at such
large distances from the Sun it is hard to understand from where the
energy to sublime signicant amounts of gas and cause a comet-like
outburst might come. An alternative is that the resurfacing might be
the result of collisions with other, smaller, trans-Neptunian objects.
The impacts might punch through the dark crust to fresh material
below it. Such an impact might produce a crater that not merely pene-
trated to bluer material, but was surrounded by an ejecta blanket of
fresh material from below the surface. Over time, which would proba-
bly be measured in millions of years, this freshly exposed material
would be irradiated and slowly darken. If this is what is happening
today, then a range of colours is fairly easy to explain the objects
most recently subjected to a few large impacts will have the most
exposed ices and so the bluest colours, the ones which have not been
struck recently will be redder.
The impact resurfacing idea was advanced in the early 1990s, but
has since been supported by work published by Susan Kern and her
collaborators in October 2000. Observations of the Centaur 8405
Asbolus (1995 GO) made with the Hubble Space Telescope in 1998
suggest that Asbolus has regions of its surface covered with very dif-
ferent materials. For technical reasons the observations of Asbolus
were made as two pointings of the satellite with a gap of over an hour
between them. Although the original intention had been to add the
two datasets together, Kern found that the two observations looked
rather different. It seemed that one region was uniformly dark and
featureless while the other had a brighter spot, perhaps with evidence
of water ice. This bright spot could be an impact crater which had bro-
kenthroughthe older crust andexposedfresh, icymaterial frombelow.
Of course, this result was unknown in 1998 and further
observations tended to cloud, rather than resolve the issue. The rst
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
119
hint of a problem came with a short but possibly important contribu-
tion from Steve Tegler and Bill Romanishin. Tegler, like many other
astronomers of his generation, had become interested in astronomy
as he followed the exploits of the Apollo astronauts exploring the
Moon in the late 1960s and early 1970s. His rst encounter with comets
came in 1973, at the age of eleven. Like many other amateur
astronomers of the time he tried and failed to nd comet Kohoutek,
which turned out to be much fainter and harder to locate than had
been predicted. He had more success with Comet Halley, studying it
with telescopes in Chile as part of a postgraduate programme which
eventually led to a PhD from Arizona State University in 1989. Whilst
in Arizona he had been introduced to the problems of CCD photome-
try by Bill Romanishin, then a postdoctoral researcher interested in
studying faint galaxies. In 1995 Tegler moved to Northern Arizona
University in Flagstaff where he once again encountered Bill
Romanishin. His old observing partner was now on the staff of the
University of Oklahoma and was visiting Arizona for the summer.
Since they had a lot of experience of photometry of faint objects, they
decided to try their luck with observations of some newly discovered
trans-Neptunians. They were able to get observing time on the Kitt
Peak 2.3 m telescope and had their rst run in November 1995. Tegler
describes their early observing runs as A struggle. Sometimes the
objects orbits were so uncertain that they could not be found in the
relatively small eld of view of their CCD camera. Sometimes the
objects were much fainter than predicted and so were impossible to
measure accurately.
Despite these difculties Tegler and Romanishin pressed on with
their programme. In 1998 they published a paper in the journal Nature
in which they suggested that the Kuiper Belt objects, and their cousins
the Centaurs, were divided into two distinct classes of basically
similar objects. Their results seemed to show that one class was
neutral in colour, or Chiron like, and the the other class was very red,
rather like Pholus. Such a bimodal distribution of colours would
imply that there were compositional differences between the two
classes, perhaps relating to the region of the protoplanetary nebula in
which they formed. Signicantly, a bimodal distribution is very
difcult to explain in terms of gradual resurfacing which clearly pre-
dicts that there should be a wide range of colours and not two nar-
rowly dened classes.
B
e
y
o
n
d

P
l
u
t
o
120
Several other groups doing similar projects did not agree with the
conclusions of Tegler and Romanishin, but worse was to come. They
could not seem to agree amongst themselves either. Part of the
problem may have been that there were few observations and the
various groups often tried to bring together data from a number of
different sources. Since each observer was using different telescopes,
and perhaps subtly different lter systems, this lack of consistency
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
121
0.2
0.4
0.6
0.8
1.0
1.2
0.4 0.6 0.8 1.0 1.2 1.4 1.6
Tegler & Romanishin 1998
KBO
Centaur
Sun
V
-
R
B-V
Figure 7.2 Two BVR colour diagrams for trans-Neptunian objects and Centaurs as pre-
sented by different groups. In (a) Jane Luu and Dave Jewitts data suggest that the objects
have a wide range of colours. (b) shows observations by Steve Tegler and Bill Romanishin
which indicate that there are two distinct groups of objects with either neutral (bottom
left) or red (top right) colours. (Dave Jewitt.)
0.2
0.4
0.6
0.8
1.0
1.2
0.4 0.6 0.8 1.0 1.2 1.4 1.6
KBO
Centaur
Sun
V
-
R
B-V
Luu and Jewitt 1996
might have produced more scatter in the data than anyone realised at
the time. Dave Tholen, who tries to do photometry with such precision
that he is sometimes known as Dr Milli-mag, speculates that part of
the problem might have been due to faint background galaxies. There
are faint galaxies all over they sky and, because they are extended,
they are quite hard to spot. However, even though they are faint, the
total light from a distant galaxy may be signicant over an area of a
few square arcseconds. So if a faint trans-Neptunian object happens
to be in front of a galaxy when it is observed, considerable galaxy light
may be included in the measurement unintentionally. The result will
be a value which is too bright. Repeating the measurement a few
hours later when the object has moved onto a different patch of sky,
which may or may not have a different galaxy in it, will produce a dif-
ferent result. The lack of agreement between different groups was a
worry, and deep down was the hint of something that few pho-
tometrists like to admit. Photometry is supposed to be a science, but
sometimes it seems to resemble a black art. Photometry of moving
23rd magnitude objects from moderate-sized telescopes presents a for-
midable observational challenge which can be approached in a
number of different ways. Perhaps some of the techniques being used
were not as good as was rst thought.
At the root of the debate was how to extract the most accurate value
for the ux of a faint Kuiper Belt object from a CCD image. A few years
ago the problem was much simpler. Most photometers used single
detectors, such as photomultiplier tubes, that viewed the sky through
a xed aperture. The detectors just counted the number of photons
which arrived in a given time. The data that was output was, in effect,
a single number per observation and that was that. The eld of view
was usually dened by a physical aperture made by drilling or etching
a hole through a piece of metal or metallic lm and putting this into
the beam of light somewhere between the telescope and the detector.
The aperture dened a circle on the sky and admitted light from only
that region. When observing, the effects of seeing and errors in the
telescope tracking spreads out the starlight, but provided the aperture
is fairly large, the detector sees almost all of this light. This of course
is good, but the large aperture means that the detector also sees con-
siderable background sky around the target. This large background
is bad because even on a dark night many photons arrive from the sky
and random uctuation in the rate at which they arrive constitutes
a source of noise in the measurement. The more sky there is in the
B
e
y
o
n
d

P
l
u
t
o
122
aperture, the larger the error in the photometry produced by the sky
noise. (The error is proportional to the square root of the number of
photons which are arriving.) Sky noise can easily swamp the signal
from very faint objects and this argues against using large apertures
for photometry of faint objects. Switching to a smaller aperture
reduces the sky noise by allowing through less background light, but
it also means that it becomes more important to ensure that the object
being measured is accurately centred in the photometers aperture. If
the target is not well centred, then some light from the target is lost.
More to the point, the amount of light which is lost varies from object
to object depending on how well each target is centred and on the
seeing at the time of the measurement. Compromises had to be made
to try and get the best ratio of signal to noise, and typically
astronomers doing photoelectric photometry used apertures which
projected 1020 arcseconds on the sky.
The advent of cameras tted with CCDs and infrared arrays revolu-
tionised photometry. Their greater sensitivity made it possible to
measure much fainter objects than before, but it also provided several
challenges in deciding how to measure them. With a traditional photo-
electric photometer, the observer made a choice of what aperture to
use, put the aperture slide into the beam and faced the consequences.
If, later on, it was realised that the original choice was not the best
one, it was just too bad. It was already too late to do anything about it.
CCD cameras opened up a huge number of new variables since each
frame recorded images containing many data points, rst thousands
and soon, as the chips got bigger, millions. Before any photometry
could be done, these images all had to be processed. Different
observers like different processing methods and use different soft-
ware packages to remove bad pixels and at eld their images. While
the various methods should all produce the same result, great care
has to be taken to ensure that, for example, the at-eld is really
at.
However, the real trouble usually starts when it comes to actually
doing the photometry. By analogy with traditional techniques, a
favoured method is to use image processing software to draw a circle
around the target and then to add up the signal in each of the encir-
cled pixels. This is analogous to measuring all the ux through a
single aperture. Of course the pixels are usually square, and circles
are round, so at the circumference of the software apertures some
allowances have to be made where the circles cut across the square
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
123
pixels. Just as in classical photoelectric photometry, the smaller the
software aperture which can be used the better, but the smaller the cir-
cular aperture becomes, the larger is the fraction of encircled pixels
which fall under the circumference of the circle. So as the aperture
gets smaller the softwares compensation for these edge effects
becomes more critical. Not only that, but since the astronomer
chooses the aperture after the observation rather than before, various
different apertures can be used on the same image. These different
apertures dont generally produce the same answer. This is because
the light from the objects under study is spread out by the telescope
optics and the seeing into a little blob on the image. Different software
apertures will include different amounts of light from the object and
background sky. Unless the images are very sharp, small apertures
tend to miss quite a bit of the light that the effects of seeing and errors
in focus have allowed to fall outside the aperture. Since the calibration
of the observations will have been made by observing stars in differ-
ent patches of sky at different times (when depending on the seeing
the star images might have been sharper, or more fuzzy, and so more
or less light might have been concentrated in the central regions of
each stellar image), care must be taken to allow for these uctuations.
A technique commonly used when measuring faint targets is to use
very small apertures. This excludes most of the sky and so maximises
the signal-to-noise ratio. However, this gain in signal to noise comes at
the expense of losing some of the light, and getting a value which in
absolute terms is too small. To correct for this, the amount of light lost
from the small aperture is then estimated by observing a much
brighter star in a series of different apertures. Since the bright star
produces images with high signal to noise in even very large aper-
tures, it is possible to trace out the amount of missing light in pro-
gressively larger apertures and calculate what is called an aperture
correction. This correction can then be applied to values from the
small apertures to allow for the missing light. Aperture correction is
another idea that sounds simple, but is actually much more difcult
than it appears. A variety of problems lurks to trap the unwary. These
include, but are not limited to, picking a reference star that is in fact
not a star at all, but is a distant galaxy and so does not have the correct
prole to start with, and having the CCD chip not exactly at right
angles to the incoming beam so the focus, and hence the aperture cor-
rection, is different at different places in the image. If this was not
enough, moving objects smear out during a long exposure so the
B
e
y
o
n
d

P
l
u
t
o
124
image of the asteroid will not have the same prole as a star in the
rst place.
Romanishin and Tegler attacked this problem very carefully. They
kept their exposures short to reduce the effects of image smearing.
They added together the same images in both the xed reference
frame of the stars, to get the best possible aperture corrections, and
then in the moving reference frame of the asteroids so they could
safely use very small apertures. The small apertures boosted the
signal-to-noise ratio and minimised the likelihood of faint galaxies
contaminating their measurements. Other groups abandoned aper-
tures altogether and used software which tted the observed prole of
the objects in the images to a shape predicted from statistics. It then
calculated the total ux using this prole. Although well suited to
faint stars, this method also runs foul of the fact that an asteroid will
move during the exposure and so its prole is not exactly what the
software is expecting.
Ultimately, there is no getting away from the fact that this is a very
tricky problem and it is hard to know which method produces the best
results. Each approach has its champions who often defend their
results with conviction. Even so, in the back of everyones mind is the
fact that there is only one right answer and sooner or later the truth is
going to come out. Fortunately, most of the individuals involved in
this debate recognise this problem and have been cooperating to try
and reach a solution. At a scientic workshop held in Germany in
November 1998

it was agreed to hold a blind test. A number of people

would try different methods on the same images and see how closely
the answers came out. Catherine (known to her friends as Cathy)
Delahodde of the European Southern Observatory in Chile agreed to
organise this test. She took some CCD frames containing a genuine
trans-Neptunian object and added 24 synthetic objects for which only
she knew the correct magnitudes. These frames were posted on an
internet site and various solar system astronomers were invited to
download them and try to reduce the data using their favourite
methods. The odd thing was, all the results came out almost the same,
with most of the groups getting values that agreed with each other to
within a few per cent. What was going on? Was it possible that the
reasons for the disagreements were not due to different techniques at
all? The observations were not all made at the same time so perhaps
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
125

Actually in a hotel bar after the end of the days ofcial presentations.
the objects themselves were varying, either due to rotation or to some
kind of comet-like activity. Both such mechanisms have been pro-
posed, and both are worth considering.
As noted in connection with the Centaurs, repeated photometry of
a solar system object may reveal that its brightness varies over time
in a manner that is not explained by changes in its distance from the
Earth and the Sun. If the changes are irregular, as was the case for
Chiron in the 1980s, then it is most probably due to a cometary out-
burst. Such outbursts can surround the object with a cloud of gas and
dust which reects sunlight and makes it appear brighter. Although in
the case of comets fairly close to the Sun the coma is usually visible
quite easily, faint comae around very distant objects may be impossi-
ble to resolve directly. An alternate approach is to compare the image
prole of the suspected comet to the prole of a star of similar bright-
ness; if the object is a comet its image will seem slightly wider and
less pointlike than a star. Given that objects beyond Neptune are gen-
erally very faint, and any extension is likely to be small, this sort of
comparison is very difcult to do from the ground. A particular
problem is the effect of seeing, which blurs out the prole of each
object and may mask any faint extension close to the suspected comet.
To get around this difculty there have been attempts to use the
Hubble Space Telescope to detect comae around Kuiper Belt objects.
The Hubble Space Telescope has very ne optics and is unaffected by
atmospheric turbulence, so it can produce very sharp images, well
suited to image prole comparisons. Although one group announced
in 1988 that they thought they had found evidence that the prole of
one Kuiper Belt object seemed to be extended, this result was not
conrmed by more careful analysis. Direct evidence for cometary out-
bursts in the Kuiper Belt thus remains lacking.
Outbursts produce irregular and unpredictable changes in bright-
ness, but rotation can cause a solar system object to wax and wane on
a regular and repeatable basis. The variation may have one of two
root causes. A spherical object which is the same colour all over, for
example a billiard cue ball, will appear the same brightness
whichever side is facing the observer, but a rotating object which has
large regions of different reectivity will vary in a regular way. At
some times the observer will see entirely the brighter side, sometimes
a bit of both and sometimes the darker face will rotate into view. This
effect will produce a regular lightcurve with one maximum and one
minimum per rotation. Of course this is only true if the different
B
e
y
o
n
d

P
l
u
t
o
126
regions are large and cover appreciable fractions of a hemisphere.
If the different regions are evenly distributed, like a soccer ball with
black spots spread uniformly over a white surface, the average bright-
ness will not change during a rotation because the different areas will
blur together and appear a uniform grey.
If the object is irregular in shape, like a potato, then as it turns it
will present different faces to the observer. Since the projected area
will change, the apparent brightness will go up when the wide face
swings into view and down again when the object is seen end on. In
this situation, the lightcurve will show a characteristic double peaked
lightcurve with two maxima and two minima (which need not be the
same) per rotation. The range of variation of such a lightcurve gives
an idea of how irregular the object is. Generally speaking, the bigger
the change the more irregular the object, although this effect varies
depending on the viewing geometry. If the object is seen close to pole
on it will have almost no lightcurve, no matter what shape it is.
The rst attempt to produce a lightcurve of a trans-Neptunian
object came from the discoverers of 1993 SC, Iwan Williams, Alan
Fitzsimmons and Donal OCeallaigh. Their inadvertent re-observa-
tion of this object during the search programme described earlier pro-
duced a total of 11 observations of 1993 SC over a period of six nights.
As the object moved slowly, it remained in more-or-less the same
stareld over the whole observing run. This made it possible to
compare its brightness with the same stars on every night of the run.
These comparisons seemed to indicate that 1993 SC was varying by
more than half a magnitude, which in turn suggested that it was quite
irregular in shape. Williams and his team used this lightcurve to esti-
mate a rotation period of about 15 hours for the object, although they
did admit that with such sparse data the result was not likely to be
very reliable. Just how unreliable it was became clear a couple of
years later when two different groups re-observed 1993 SC and found
almost no variation at all. It now seems that Williams group was
misled by unsuspected errors in the photometry extracted from their
images. In retrospect, perhaps these problems should have sounded a
warning that photometry of these objects was harder than it seemed.
Despite the difculties of trying to measure lightcurves, a few
people have persisted in the effort. First to publish some detailed
results were Bill Romanishin and Steve Tegler, who used the same
dataset on which they had based their conclusions about the colours
of distant asteroids. They found that many of these objects had no
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
127
detectable lightcurves and in several cases the maximum possible
variations were only a few per cent, the likely precision of their mea-
surements. Unlike similar studies of brighter objects, for which it is
possible to build up a lightcurve containing many points during a
single rotation, the data for the Kuiper Belt objects was generally
rather sparse. Even so, in a few cases, it did suggest that there were
signicant variations over periods of a few hours. If these changes
were due to the rotation of a spherical object with hemispheres of
very different reectivities, the resulting single peaked lightcurve
would imply rotation periods of only a few hours. This is so rapid that
the objects would be hard put to stay together; the rotation speed at the
equator would be rapid enough to throw material off the surface and
into space. Accordingly, Romanishin and Tegler interpreted the data
as double peaked lightcurves from irregularly shaped objects. This
gave rotation periods of 610 hours, similar to main belt asteroids in
the same size range, and more physically realistic.
Romanishin and Tegler took their conclusions one step further.
They noted that the objects with measurable lightcurves were the
intrinsically faintest ones and that the objects which did not vary
were generally brighter. They suggested that this was because the
larger objects are massive enough that their gravity causes them to
collapse into a sphere while the fainter, and so smaller ones, cannot
collapse and so are irregular in shape. Based on an assumed reectiv-
ity of 4% for a typical Kuiper Belt object, they estimated that the tran-
sition size from spherical to irregular was a diameter of about 250 km.
Other groups have tried to follow up this work, but it remains very
difcult. Alan Fitzsimmons group observed two objects, 1996 TO
66
and
1994 VK
8
in 1997. Although they did detect what seemed to have been
signicant changes in the brightness of 1994 VK
8
, they were unable to
conrm any particular rotation period from their data. They found no
detectable lightcurve for 1996 TO
66
. A rather more interesting conclu-
sion about 1996 TO
66
came from a European team. Olivier Hainaut,
Cathy Delahodde, Antonietta Barucci and Elisabetta Dotto had deter-
mined a lightcurve for 1996 TO
66
based on observations taken in 1997
and 1998 from Chile. These observations showed an almost symmetri-
cal double peaked lightcurve indicating a shape-dominated rotation
period of 6.25 hours. The peak-to-peak amplitude variation during the
rotation was found to be 0.12 magnitudes. Just as they were preparing
their result for publication, they learned that 1996 TO
66
had been
observed in September 1998 from Mauna Kea by Karen Meech and
B
e
y
o
n
d

P
l
u
t
o
128
James Gerbs Bauer. Olivier Hainaut had been a research fellow in
Hawaii before moving to Chile and had worked with Karen Meech
before, so they agreed to combine the data from all three runs.
Interestingly, the new observations were not consistent with the
lightcurve derived from the two previous years. The 1998 observations
suggested that 1996 TO
66
had a single peaked lightcurve with a much
larger amplitude of 0.33 magnitudes, although the rotation period
seemed to be exactly the same.
Hainaut and his co-workers examined several possible explana-
tions for this and concluded that the best one was a brief phase of
cometary activity occurring sometime between the 1997 and 1998
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
129
Figure 7.3 Top. An image of trans-Neptunian object 1996 TO
66
. Bottom. The light curve
deduced from observations made at the European Southern Observatory. (ESO.)
observations. They suggested that during the outburst a region of
the objects surface was modied, probably by the deposition of fresh
bright ices which covered up some otherwise darker material. Such
activity might, indeed probably would, also change the observed
colour of the object, perhaps explaining some of the colour differ-
ences reported by various groups. Further observations of the same
object were made in 1999 using the 8.2 m European Very Large
Telescope in Chile. This data implied a similar rotation period, but the
amplitude of the lightcurve had changed again. This time the varia-
tion was about 0.2 magnitudes and the shape of the curve was differ-
ent. According to Hainaut, this could conrm the hypothesis that the
object is being actively resurfaced on short timescales.
For the moment, many of these questions about lightcurves and
surface colours remain unresolved. The observations are difcult, the
signal-to-noise ratio is often low and the uncertainties in the data may
be larger than people are yet willing to admit. It will be a little while
before we know the details for certain, but what is already clear is that
the physical properties of the objects in the Kuiper Belt are at least as
complicated as the orbital dynamics. Whatever the nal resolution of
the debate over the colours and rotation rates, photometry is unlikely
to solve the problems of the chemical composition or mineralogy of
the outer solar system. Two-colour diagrams are useful for sorting
objects into broad groups, which may have roughly similar surfaces,
but the lter passbands cover a wide range of wavelengths. Filter pho-
tometry may mask a variety of spectral features which could be used
to identify specic minerals, or molecules, if only the objects could be
observed with much higher spectral resolution.
Although Dave Jewitt and Jane Luu tried to obtain an optical spec-
trum of the relatively bright Plutino 1993 SC as early as 1994, spec-
troscopy of Kuiper Belt objects, especially in the near infrared region
where ices have identiable spectral features, is not really a practical
proposition with 24 m class telescopes. It was not until solar system
astronomers got their hands on bigger mirrors that much progress
could be made in understanding the composition of these distant
objects. The rst results had to wait for spectroscopy from the huge
Keck telescopes on Mauna Kea.
The W. M. Keck Observatory comprises two of the largest optical
telescopes in the world and a headquarters building in the sleepy
cattle town of Waimea, on the Big Island of Hawaii. The observatory
was nanced by a donation from the W.M. Keck Foundation to a con-
B
e
y
o
n
d

P
l
u
t
o
130
sortium comprising the California Institute of Technology and the
University of California. The objective was to build a new large tele-
scope to take advantage of the clear skies available over Mauna Kea.
Since, at the time the project was conceived, making very large single
mirrors was known to be very difcult, the Keck telescopes are of an
innovative design. They comprise numerous hexagonal mirrors, each
about 2 m in diameter, tted together to make a single collecting area
equal to a mirror 10 m across. The whole array of hexagonal mirrors
is supported by an ingenious backing structure which keeps them all
pointing in the same direction so they function as a single mirror. So
successful was this concept that once the rst telescope was opera-
tional, the Keck foundation was persuaded to provide funds to build a
second telescope, increasing the total sum they had donated to about
$140 million.
Since the Keck Observatory was owned by the Californian consor-
tium, all the observing time, apart from the fraction claimed by the
IfA in Honolulu for allowing the telescopes to be built on Mauna Kea,
originally belonged to these two establishments. However, from
October 1996 NASA entered into an agreement to provide some of the
running costs of the Keck Observatory in return for one third of the
observing time on one of the two telescopes. NASA, which already
funded the much smaller IRTF telescope on Mauna Kea, needed access
to a large telescope for its wide-ranging programme for studying the
origins of planetary systems. Contributing to the Keck Observatory
was the quickest and most cost-effective way of doing this.
In early October 1996, some of this observing time was allocated to
a group including the University of Arizonas Robert (Bob) Brown and
NASAs Dale Cruikshank. They used it to take infrared spectra of the
Plutino 1993 SC with the Keck observatorys dual-purpose near-
infrared instrument NIRC. Although NIRC stands for Near InfraRed
Camera, the instrument can also be used as a low resolution spectro-
graph. Spectroscopy is done by inserting some extra optics into the
cameras optical path and using a grism (a combination of grating and
prism) to disperse the incoming light into a spectrum. For this to
work, the grism mode includes a narrow slit which is projected onto
the sky and which blocks most of the eld of view. Only light passing
though the slit reaches the grism and is dispersed into a spectrum.
Accordingly, the image of the object under observation has to be
placed exactly in the slit and kept there throughout the observation.
Luckily, the fact that NIRC also works as a camera makes this exercise
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
131
of nding a faint target and getting it lined up with the slit slightly
easier than it would be for a pure spectrograph. It is possible to switch
NIRC between the imaging and spectroscopy modes to see exactly
where the target appears in the instruments eld of view. Brown and
colleagues started their observations by pointing at the predicted
position of 1993 SC and taking a couple of images using the instru-
ment in its camera mode. By comparing these two pictures, they
quickly identied which object was their target and moved the tele-
scope to place 1993 SC in the correct place to send its light through the
slit. Once this was done they switched to spectroscopy mode. As is
usual with spectroscopic observations, they also observed some stars
in order to allow for the effects of the Earths atmosphere, which does
not transmit equally well at all wavelengths. After calibrating their
data and using data from their comparison stars to remove any fea-
tures due to the atmosphere, they had the rst infrared spectrum of
a Kuiper Belt object. What did it mean?
The spectrum was fairly noisy, which is not surprising since their
target was so faint, so they smoothed it by combining some of the
points to improve the signal-to-noise ratio at the expense of lowering
the spectral resolution. The result was an interestingly complex spec-
trum that seemed to show a number of spectral features. Although the
resolution of the spectrum was not good enough to identify the fea-
tures unambiguously, the overall shape was rather similar to spectra
of Pluto and of Neptunes largest moon Triton. Spectra of Pluto and
Triton taken during the 1980s suggested that their surfaces might
include some methane ice trapped in a solid solution of frozen nitro-
gen, and while the spectrum of 1993 SC did not prove that the exact
same situation existed there, it was certainly suggestive of a surface
containing a frozen mixture of light hydrocarbons. The existence of
solid nitrogen on 1993 SC was a bit more problematical, since given
the likely temperature of the surface (about 50 K, 50 degrees above
absolute zero or 223 degrees Celsius) and its low gravity, it is not
certain that solid nitrogen could survive there over the age of the
solar system.
A second spectrum of a trans-Neptunian object was not long in
coming. Dave Jewitt and Jane Luu observed the scattered disc object
1996 TL
66
from two different telescopes in September 1997. They used
the NIRC camera/spectrometer on one of the Keck telescopes to take
an infrared spectrum and combined this with an optical spectrum
taken at the Multiple Mirror Telescope, or MMT, located on Mt
B
e
y
o
n
d

P
l
u
t
o
132
Graham in Arizona. The MMT is another telescope of unusual design.
In its original form it combined the light from ve mirrors at a single
focus to produce the light-gathering power of a single 4.5 m mirror.
However, unlike the Keck telescopes, these mirrors were not hexagons
tiled together into a single surface, but conventional circular mirrors
mounted in a common framework. The design worked, but the tele-
scope was closed in 1998 to remove the complex multi-mirror system
and replace it with a single mirror 6.5 m in diameter.

Jewitt and Luu nished up with three spectra, one from the MMT
covering the optical region and two from the Keck telescope. Between
them these three spectra encompassed the entire optical and near-
infrared region. They carefully combined the different spectra, using
photometric measurements taken in both the optical and infrared
regions to ensure that they did this correctly. The nal result was com-
pletely featureless. The spectrum showed no sign of the infrared fea-
tures which Bob Brown and his group had found in 1993 SC. There was
no evidence for water ice or any hydrocarbons. All in all, the spectrum
of 1996 TL
66
was rather dull and featureless, just like the spectra of
some of the Centaurs.
A third spectrum of a trans-Neptunian object was published in
1999. Once again, this came from Bob Brown and his collaborators
using NIRC on one of the Keck telescopes. This time the target was
1996 TO
66
and the result resembled neither 1993 SC nor 1996 TL
66
. The
spectra, which were taken and reduced in a similar manner to their
earlier data, showed two absorptions in the 12 micron region. These
features suggested that there was water ice on the surface of 1996
TO
66
. Not only this, but spectra taken on two successive nights showed
that the amount of water, as determined by the depth of absorption
features in the spectra, was variable. Here, at last, was evidence that
the Kuiper Belt objects might have global variations in their surface
properties. Perhaps the surface of 1996 TO
66
was patchy, with some
regions containing more fresh ice than others, as suggested by the
changing lightcurve reported by Olivier Hainaut and his colleagues.
The likely presence of ice on 1996 TO
66
was conrmed by observa-
tions made in 1998. During August and September, Keith Noll, Jane
Luu and Diane Gilmore attempted to observe ve trans-Neptunian
objects using the NICMOS near-infrared photometer/spectrometer
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
133

Despite this change, there are no plans to rename the telescope, it will still be known as
the MMT.
F
i
g
u
r
e

7
.
4
S
p
e
c
t
r
a

o
f
t
w
o

t
r
a
n
s
-
N
e
p
t
u
n
i
a
n

o
b
j
e
c
t
s
.
A
t

t
h
e

t
o
p

i
s

t
h
e

s
p
e
c
t
r
u
m

o
f
1
9
9
3

S
C

t
a
k
e
n

b
y

B
o
b

B
r
o
w
n

a
t

t
h
e

K
e
c
k

T
e
l
e
s
c
o
p
e
.
B
e
l
o
w

t
h
i
s

i
s
t
h
e

c
o
m
b
i
n
e
d

M
M
T
/
K
e
c
k

s
p
e
c
t
r
u
m

o
f
1
9
9
6

T
L
6
6
t
a
k
e
n

b
y

D
a
v
e

J
e
w
i
t
t

a
n
d

J
a
n
e

L
u
u
.
(
D
a
v
e

J
e
w
i
t
t
.
)

on the Hubble Space Telescope. The objects were too faint for spec-
troscopy, but in each case they planned to take measurements through
a series of six lters covering the 12 micron range. Combining these
would give a rough spectral ngerprint of each object and should
reveal any signicant spectral features. One observation missed its
target. The satellite had been pointed at the position requested, but
the object was not to be seen. This was because the best orbital ele-
ments then available for the object were not correct, causing the tele-
scope to be pointed at the wrong position. The other four observations
all went well. After applying the usual at eld corrections and
removing cosmic rays from the images, Noll and his co-workers
extracted the best photometry they could from their data. They then
combined their infrared data with some optical observations from the
literature and looked at the results. Their data showed that 1996 TO
66
was rather different from the other three objects which they had
observed. 1996 TP
66
, 1996 TQ
66
and 1996 TS
66
all got redder at longer
wavelengths and had a few dips that hinted at some as yet unknown
spectral features. In some ways these three objects were broadly
similar to the redder Centaurs. On the other hand, 1996 TO
66
was
bluishand the photometry throughthe various infrared lters showed
that it had dips inits spectrum. These dips were consistent withBob
Browns 1999 Keck spectrumshowing water ice onthe surface.
The presence of ice on 1996 TO
66
was further conrmed by Mike
Brown, of the California Institute of Technology. As a member of the
Caltech faculty Brown has access to the Keck telescopes and he had
used them to survey six Kuiper Belt objects. However, no ice was seen
in the spectra of any of the other ve objects. All appeared featureless,
which Brown says he found interesting. Like many other people he
had expected the objects to have spectral features typical of those seen
on the red Centaur Pholus. After all, if Pholus looks the way it does
because it has only recently left the Kuiper Belt, then the objects still
there ought to look just like it. The fact they do not casts doubt on the
commonly accepted view that Pholus has a very old surface.
Mike Brown nds his results strange, and wonders if the Classical
Kuiper Belt might contain objects from two distinct populations.
However, Noll, Luu and Gilmore sounded a warning that this
dichotomy did not, in itself, support the case for a bimodal distribu-
tion of trans-Neptunian objects. While it might appear red colours
would correspond to objects dominated by organic rich solids, and
blue ones to objects dominated by water ice, the situation was clearly
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
135
more complex. The red Centaur Pholus has water ice features in its
spectrum, but the bluish trans-Neptunian 1996 TL
66
does not show any
evidence for ice.
Supporting what might be described as this conclusion of confu-
sion are a number of ground-based observations. Dave Jewitt and
Jane Luu observed ve objects in the infrared JHK bands using one of
the Keck Telescopes. They combined their data with visible photome-
try taken from the University of Hawaii 2.24 m telescope a few weeks
later to determine visibleinfrared colours for ve objects. Soon after,
John Davies, Simon Green, Neil McBride, Dave Tholen and their stu-
dents went one step further. They used the University of Hawaii
2.24 m and the nearby UKIRT telescopes to observe a dozen or so
objects in the visible and near infrared simultaneously. By observing
in this way, it was possible to remove any uncertainties in the colours
which might result from the unknown effects of lightcurves or hypo-
thetical cometary outbursts. This observing strategy involved much
telephoning from dome to dome throughout the night in order to syn-
chronise the observations. Both groups obtained sets of
visibleinfrared data which indicated that there was a wide range of
colours that were not correlated with anything in particular.
However, Steve Tegler andBill Romanishinsoonreturnedtothe
fraywithanewresult. The controversysurroundingtheir initial
announcement that the trans-Neptunianobjects fell intotwowell-
denedcolour groupings hadenabledthemtoget time onthe Kecktele-
scopes tocontinue their programme. In1998 and1999 theyobserved17
more objects fromthe Kecktelescopes andinAugust 2000 publisheda
paper inNature defendingtheir original conclusions. Theyalsonoted
that there seemedtobe acorrelationof colour withdistance fromthe
Sun. Theyreportedthat objects innear-circular orbits beyondabout 40
AUfromthe Sunwere consistentlyveryred, while objects closer in
couldfall intoeither of their neutral or redclasses.
Tegler and Romanishins new data were presented in the form of a
talk at the October 2000 meeting of the American Astronomical
Societys Division for Planetary Sciences in Pasadena, California.
Immediately after it was nished, Dave Jewitt rose to present his
results. Jewitt said rmly that his data, which also included observa-
tions from the Keck telescope, showed neither a bimodal colour distri-
bution nor any trends of colours with distance from the Sun. Perhaps
as the organisers had intended, it was in direct contradiction of the
talk which had just nished.
B
e
y
o
n
d

P
l
u
t
o
136
There was no resolution of the issue at the meeting. Later one of
the audience remarked that the bimodal distribution seemed to get
less convincing as more data was presented, which is the opposite of
what would happen if the effect was real, but Steve Tegler says he is
sticking to his conclusions. Antonietta Barucci and her colleagues
have continued their observing programme and are in the other camp;
like Dave Jewitt they believe that their data do not support a bimodal
distribution nor any trends with distance from the Sun. The issue is
clearly a complicated and controversial one, and it is going to take
some time and effort to unravel it.
W
h
a
t

a
r
e

l
i
t
t
l
e

p
l
a
n
e
t
s

m
a
d
e

o
f
?
137
Numbers
and sizes
To take full advantage of what the trans-Neptunian region can reveal
about the formation of the solar system, astronomers need to under-
stand more than just the physical characteristics of a few dozen
objects. They also require details of the size distribution of the popu-
lation to compare with their models of how the planets formed. How
much material is there beyond Neptune? Is this consistent with obser-
vations of other stars thought to be forming planets?
Size distributions are generally suchthat there are many small
objects for eachlarge one. They are usually expressed interms of the
number of objects larger thana certainsize plotted against that size.
Since the number of objects increases rapidly as the size goes down,
and the graphmust represent a wide range of sizes froma fewnear-
planet-sized objects to anuncountable number of dust grains, both
axes are normally plotted onlogarithmic scales. The result is usually
a more-or-less straight line whose slope is called the power lawindex.
The power lawindex for a populationwhichcontains objects which
are still growing is different fromthat of a populationwhichhas
stopped evolving, or one whichis inthe process of grinding itself back
downinto smaller pieces. So, the size distributionis critical to under-
standing the evolutionary history of the trans-Neptunianregion.
Determining the size distribution of objects in a laboratory is just a
matter of measuring and counting the sample. Attempting to do the
same thing for a population of faint astronomical objects, most of
which have not even been discovered yet, is rather more problemati-
cal. One of the difculties is the issue of selection effects. Just how
representative of the population as a whole are the objects discovered
so far? While there is no need to discover every last object before
trying to determine a reliable size distribution, it is important to
C
H
A
P
T
E
R

8
139
make sure that the sample which has been observed is a fair selection
of those that remain to be found. Samples of astronomical objects are
notoriously loaded with observational biases and many people have
found that failure to take these effects into account in statistical
studies can result in seriously awed conclusions.
As an example of an observational bias, consider that before they
can be included in any census of the trans-Neptunian region, the
objects must be discovered. Broadly speaking, the likelihood of a solar
system body being discovered depends on its brightness. The brighter
it is the more likely it is to be found using a medium-sized telescope,
and there are a lot more medium-sized telescopes around the world
than there are really big ones. At any given time the brightness of a
solar system object depends on its size, its reectivity and, since it
shines by reected light, on its distance from both the Earth and the
Sun.

Since the trans-Neptunian region begins at about 40 AU, and the

Earths orbit is by denition at 1 AU, the precise observing geometry
does not make much difference to the distances involved, but it is clear
that for a given size and reectivity, the further an object is from the
Sun the fainter it will be. The fainter it is, the harder it is going to be to
discover. This effect is seen very easily in the statistics of the known
trans-Neptunian objects. About one third of them are Plutinos, but
this does not necessarily mean that the Plutinos are this common. It
just shows that if all other things are equal Plutinos, being on average
closer to the Sun, are easier to discover.
There is another more subtle factor which mitigates in favour
of the Plutinos. To make it into the list of known trans-Neptunian
objects, an object has to be not just discovered, but re-observed on sub-
sequent months. Only then can its orbit be determined. Since the
observed arcs of most newly discovered distant objects are small,
there is often not enough information to do much more than make a
good guess at some of the orbital elements. Consequently, projecting
an objects position forward so it can be reobserved a few weeks or
months later often requires the staff of the Minor Planet Center to
make a few assumptions. One of the assumptions they may make is
that the object is in an orbit stabilised by a resonance, such as the 2:3
mean motion resonance with Neptune which marks an object as a
Plutino. If this assumption is correct, then the object will duly appear
B
e
y
o
n
d

P
l
u
t
o
140

There are a number of other subtle factors which affect the observed brightness of solid,
rough, solar system objects and which depend on the exact angle between the Sun, the
object and the Earth, but these do not concern us here.
close to the prediction, be recovered and added to the catalogue of
known minor planets. However, if this initial guess is wrong, then the
calculated future positions may be in considerable error and the
object may not be where the observers trying to recover it are looking.
If this happens then it is likely that the object will not be recovered.
Unless it is discovered again by chance, it will join the legion of lost
trans-Neptunians. What may have been an early example of this was
the fate of 1993 RP, one of the rst trans-Neptunian objects discovered.
Brian Marsden assumed that, like the other objects discovered that
autumn, it was in the 2:3 resonance. He made his predictions accord-
ingly, but 1993 RP was never recovered. Perhaps it was not a Plutino,
but something else, for example a scattered disc object. Unless 1993 RP
is found again by chance, we will never know. The result of this sort of
bias is that a higher fraction of Plutinos and other objects in orbits
stabilised by likely resonances are recovered and they swell the list
of known objects in a disproportionate way.
Additional observational biases can arise from the way in which
the sky is searched. We saw earlier that during their long-running
search programme, Dave Jewitt and Jane Luu concentrated on
regions of the sky where the ecliptic plane was well separated from
the Milky Way. They, and others, did this to increase their chances of
nding faint moving objects by restricting the number of stars in each
of their images. However, the use of such observing strategies means
that some regions of the sky have not been searched, or have been
searched with a lower probability of success, than others. If the trans-
Neptunian objects are distributed randomly around the Sun, then this
will not matter; one patch of sky will be as good as any other. However,
if, as seems likely, certain stable regions contain a disproportionate
number of objects, the resulting discoveries will be biased depending
on whether or not the populated regions happen to lie in the same part
of the sky as the galactic plane.
These sorts of observational biases can be estimated and allowed
for, but there still remains the problem of actually calculating the
sizes of the objects that are being counted. If all that is available is a
measurement of how much sunlight is reected back from an object,
its size can only be determined if its reectivity is known. Without
this information it is impossible to tell if one is looking at a small
bright object (such as a snowball) or a large dark object (such as a
lump of black rock). The reectivity of planetary surfaces is
described in terms of albedo and is 1 (or 100%) for a perfect reector
N
u
m
b
e
r
s

a
n
d

s
i
z
e
s
141
and 0 for something which is truly and utterly black. Most asteroids
have albedoes in the range 0.04 (4%) to 0.16 (16%). While this range
might not sound very large it can produce a variation in the estimated
diameter of an object by a factor of two and this translates to an error
in the mass of a factor of eight. Inactive comet nuclei have albedoes of
210% while Pluto, which has a surface covered at least in part by ice,
has an albedo approaching 60%. Clearly, before sensible size estimates
can be made for the trans-Neptunian objects, astronomers need to
decide which albedo value they are going to use for their calculations.
The albedo of an asteroid can be found if simultaneous measure-
ments are made of both the visible light being reected back from its
surface and the amount of heat which it is emitting. This determina-
tion is possible because, provided the object is in thermal equilibrium
with the incoming sunlight, the sunlight that is not reected must be
absorbed. The energy absorbed goes into warming up the asteroid and
is then radiated back into space. The peak wavelength of this thermal
emission is determined by the objects temperature; the hotter it is the
shorter is the wavelength at which most of the emission occurs.
Provided that both the reected and emitted energy can be measured
at the same time, it is possible to use models of the objects likely
thermal properties to determine both its size and albedo unambigu-
ously. This technique has been applied to determine the sizes of
objects both close to the Earth and in the main asteroid belt. In the
inner solar system asteroids have temperatures of between 200 and
300 K and their strongest thermal emission is at wavelengths of about
10 microns. This is a region to which the Earths atmosphere is fairly
transparent and so it can be studied from ground-based telescopes.
Objects beyond Neptune are rather colder than this, typically 60 to
80 K. At these low temperatures the bulk of the thermal emission is at
longer wavelengths which do not penetrate the Earths atmosphere
and so cannot be measured from the ground.
One stepping stone to the solution of this difculty is to observe the
larger Centaurs, which are closer to the Sun and so are both warmer
and brighter. If the Centaurs are objects evolving inwards from the
trans-Neptunian region, then their albedoes should be representative
of their more distant cousins. The brighter Centaurs have signicant
thermal emission in the 20 micron region, where there is another
atmospheric window for ground-based observations, and thermal
infrared observations have been made for Chiron, Pholus and
Chariklo. The early observations of Chiron and Pholus were made by
B
e
y
o
n
d

P
l
u
t
o
142
a number of people, including Humberto Campins, John Spencer,
John Davies and Mark Sykes. They used simple infrared photometers
on telescopes such as the IRTF and UKIRT on Mauna Kea where the
thin, dry atmosphere is well suited to work in the thermal infrared.
Observations of Chariklo were made, also from UKIRT, by Dave Jewitt
and Paul Kalas, using a mid-infrared camera called MAX. Jewitt and
Kalas obtained just one data point, a measurement of the 20 micron
ux, for Chariklo which they quickly published as a letter in the
Astronomical Journal. When challenged that writing a whole paper
around a single data point was a bit excessive, Jewitt was unabashed.
Ah yes, he said, but its a very good data point. Jewitt, Kalas and the
others agree that combining the infrared data with suitable asteroid
thermal models leads to the conclusion that the Centaurs Pholus and
Chariklo are very dark, with albedos of around 4%. The albedo of
Chiron is rather harder to determine due to the likely presence of
coma, which confuses the measurements. It is thought to be a little
higher than the other Centaurs, perhaps as much as 10%.
Attempts by a group led by the Max Planck Institute for Aeronomie
in Germany to measure the thermal radiation from trans-Neptunian
objects using the orbiting European Infrared Space Observatory (ISO)
proved very difcult. Although the satellite could observe at wave-
lengths between 60 and 100 microns, where the objects were expected
to be brightest, the ISO telescope had only a small mirror with limited
light-collecting power. The telescope also had fairly low spatial resolu-
tion at these long wavelengths. Compounding the problems was that
the basic long-wavelength photometer turned out to be less sensitive
than predicted and so the project had to go ahead with a different
instrument mode. This amounted to a camera, comprising a three by
three array of very large pixels, observing through a lter centred at
a wavelength of 90 microns. The low spatial resolution that resulted
meant that it was difcult to determine how much infrared radiation
was coming from the object of interest and how much was coming
from the background. This further complicated what were already
difcult observations.
Unwanted background radiation could come from a number of
sources. These included one or more distant galaxies which just hap-
pened to fall in the eld of view, clouds of interstellar dust within our
own galaxy and the warm zodiacal dust inside the solar system. To
remove any signal from distant background sources, ISO observed the
same region twice, once when the trans-Neptunian object was in it
N
u
m
b
e
r
s

a
n
d

s
i
z
e
s
143
and again when it had moved away. By subtracting the two observa-
tions, all that remained ought to have been the ux from the target
object. However, this observational technique created an additional
complication. The brightness of the zodiacal dust depends on the
viewing geometry and, since the Earth had moved between the two
observations, the second time around the satellite was not looking
through exactly the same column of dust. There was also a possibility,
not all that remote as it turned out, that a small main belt asteroid
might appear in the beam by chance, further contaminating the data
on the distant object. As an insurance against this possibility, each
observation was taken in a sequence which included images with
ISOs short wavelength camera ISOCAM. The ISOCAM images were
taken at a wavelength of 10 microns, where any trans-Neptunian
object would be invisible, but a closer and warmer, main belt asteroid
would be quite easy to see. It was a wise precaution; one such asteroid,
a newly discovered 1 km sized object designated 1997 SU
15
, did indeed
appear to be in the beam during one of the observations.
Despite all these hurdles, ISOdoes appear tohave detectedat least
twotrans-Neptunians, albeit withrather lowsignal-to-noise ratios.
The Plutino1993 SC, together withits backgroundreference elds, was
observedtwice. The twosets of observations were made about ayear
apart. The scattereddisc object 1996 TL
66
andits reference eldwere
observedonce. The 1993 SCobservations were fairlysuccessful, getting
roughlythe same answer eachtime. Theysuggestedthat the object had
adiameter of about 300 kmandanalbedoof around23%. Interpreting
the 1996 TL
66
observationwas not sostraightforward. Althougha
source withabout the expecteduxwas detected, it appearedinthe
wrongone of the nine pixels inthe cameras eldof view. Noamount
of trying, andthe teamdidtry, couldexplainwhythe source was not in
the central pixel. Fromthese limitedobservations it was not possible
to determine the size andalbedoof 1996 TL
66
withanycertainty, but the
fact that somethingwhichwas probablythe object was detectedat all
hintedthat it was indeeddark. This conclusionfollows fromthe fact
that if 1996 TL
66
hadahighalbedo, thenits optical brightness would
have impliedthat it was small, andasmall 1996 TL
66
wouldnot have
emittedenoughinfraredradiationtohave beendetectable byISOat all!
The conuence of the three lines of evidence from comet nuclei,
Centaurs and the ISO measurements, all point to albedoes of around
4% for the trans-Neptunian objects. So, in the absence of anything
better, this value is generally assumed to be about right. However,
B
e
y
o
n
d

P
l
u
t
o
144
before moving on, let us just sound one note of caution; all the trans-
Neptunians may not be the same. As we have seen, the results from
optical and infrared photometry suggest that these distant objects
have a range of colours and spectroscopic observations show the pres-
ences of ices on some, but not all of them. Ices can be very reective
and if 1996 TO
66
has an icy surface, then its estimated diameter of
750 km might have to be revised downwards, perhaps closer to 300 km.
So the adoption of a constant value for the albedo of what may be a
very diverse population may be an unwarranted assumption.
However, until more denitive data are available, it will have to do
for the moment.
In order to tackle the question of the size distribution, various
groups have set out to try and measure the cumulative luminosity
function of the trans-Neptunian objects. This function, the number of
of objects per unit area of sky which are brighter than some specied
magnitude, is now fairly well understood for V magnitudes between
about 20 and 26. The problems arise at the extreme ends of this distrib-
ution when trying to assess the numbers of quite bright, and of very
faint objects.
At the bright end there is just one object, Pluto, which shines at a V
magnitude of about 14. This is 250 times brighter than the next bright-
est object in this region of space. Pluto was found by Tombaugh using
1930s technology and while few people doubt that he did not miss any
other bright objects, it is certainly possible that a few fainter ones
might have slipped past his tired eyes from time to time. Since there
are so many variables in trying to assess just how thorough
Tombaughs search was, and since Pluto may be unusual in a number
of other respects (for example it has a satellite and an atmosphere),
it is not always included in determinations of the slope of the trans-
Neptunian luminosity function. The outer solar system search con-
ducted by Charles Kowal, during which Chiron was discovered, is
another survey which ought to be useful in setting limits on the
number of bright trans-Neptunians. Kowal covered a fairly large
region of sky without nding any very-distant objects, but once again
it was a manual search done by blinking photographic plates. Since
even slow-moving objects smear out during long exposures, making
them that much harder to nd, it is also difcult to quantify exactly
how sensitive Kowals search was.
At the other end of the scale, Anita Cochrans attempt to determine
the number density of very-faint objects has also resulted in a fair
N
u
m
b
e
r
s

a
n
d

s
i
z
e
s
145
degree of controversy. Cochrans attempts to discover the Kuiper Belt
from the McDonald observatory in the early 1990s were foiled by a
combination of bad weather and poor seeing. These conditions meant
that her programme could not compete with searches being done from
better observing sites such as Mauna Kea. Her solution was to go to a
telescope where the image quality was guaranteed to be excellent and
weather was not going to be a problem. Anita Cochran took the search
into space, winning time on NASAs Hubble Space Telescope to search
for comet-sized objects beyond Neptune. Her objective was to conrm
the link between the Kuiper Belt and short-period comets. She was
joined in this effort by Martin Duncan, Hal Levison and Alan Stern.
Thirty four times in August 1994 the Hubble Space Telescope aimed
itself at an otherwise undistinguished eld in the constellation of
Taurus. Especially chosen since it contained few stars and galaxies,
the eld lay on the ecliptic plane, a prime hunting ground for trans-
Neptunians. Unlike most other searches from ground-based tele-
scopes, the eld chosen did not lie close to the opposition point.
Instead, the telescope was pointed at quadrature, at right angles to
the EarthSun line. Along this direction the Earth is moving almost
directly towards the line of sight and so there is no reex motion
caused as the Earth overtakes distant objects. Instead, any apparent
movement of objects seen in the eld of view is due to their actual
orbital motion. For an object beyond Neptune, this amounts to a little
less than 1 arcsecond per hour.
Inprinciple, movingobjects couldbe identiedbylookingfor the
tinystreaks theyformedas theydriftedacross the cameras eldof
view. Inpractice, since the telescope orbits above the atmosphere, and
sois completelyunprotectedfromcosmic rays, images fromthe HST
cameraare litteredwithhundreds of streaks, blobs andhot pixels
causedbycosmic rays strikingits detectors. The normal technique for
removingthese defects is tocombine manyimages andrelyonthe fact
that the real objects staystill relative toeachother, while the cosmic
rayhits appear at randompositions across the frame. Unfortunately,
while the standardprocessingtechniques wouldcertainlyhave
removedthe badpixels fromCochrans images, theywouldjust as cer-
tainlyhave removedall the objects of interest as well. Instead,
Cochrans groupadoptedatwo-stage reductionprocess. First, they
made anormal image of the eldwhichremovedthe badpixels and
retainedall the xedstars andgalaxies. Thentheysubtractedthis nal
image fromall 34 of the original frames, leavingthemwithaset of
B
e
y
o
n
d

P
l
u
t
o
146
frames containingjust badpixels, noise and(hopefully) some distant
objects. Next theyaddedtogether all of the frames usingvarious rates
of motiontypical for objects intrans-Neptunianorbits. If anysuch
objects were inthe eldof view, theninthe shiftedandsummedimages
theywouldalways landonthe same pixel andappear as discrete bright
spots. The noise andbadpixels wouldbe spreadout over numerous
pixels andbe dilutedintoinvisibility. After makinganumber of such
images, using154 slightlydifferent rates of motion, Cochrans group
examinedthe results, but foundnoobvious candidate objects.
Undeterred, they developed an automatic routine which looked
for groups of pixels that seemed a bit brighter than the background
around them. They then noted these positions as being worthy of
further study. Of course, since the objects were only just above the
background, it was still possible that they might have been chance
alignments of noise. To try and eliminate such false detections, they
tried splitting the data up into different sets (all the odd numbered
frames, all the even numbered frames, the rst 17, the last 17 and so
on). They only considered as real those detections which occurred in
the same pixel in images made with a number of different combina-
tions of frames. They also tried making images that were shifted in
ways that were not consistent with the motion of real Kuiper Belt
objects. They searched these images for faint sources in the same way.
Since the only thing that was different between the real images and
the false ones was the assumed motion of any objects in the elds,
these control frames should have provided a measure of the likely
number of false alarms caused by random coincidences of noise
spikes and other artifacts.
The result of this heroic effort in image processing was that they
detected 53 sources in the images processed using realistic rates of
motion and only 24 in the ones using demonstrably false rates. Their
conclusion was that they had indeed detected a population of small
objects at about the distance of Pluto. Assuming the albedo of the
objects was the oft quoted 4%, then, with visual magnitudes of around
28, their objects were between 5 and 10 km in diameter. This was
exactly what was expected for a population of comet nuclei waiting to
be sent towards the Sun by gravitational perturbations. Of course it
was not possible to say which of the 53 sources were real comets and
which were noise, nor was it possible to use the observations to deter-
mine the objects orbital elements so they could be followed up later.
The result was a statistical detection. There were more sources
N
u
m
b
e
r
s

a
n
d

s
i
z
e
s
147
moving in the right sort of way than there were going in the wrong
way, so some of them must be real. Or so it seemed.
Doubts about the reliability of Cochrans result soon began to
surface since it implied that small objects were tens, if not hundreds,
of times more common than would be expected based on the number
of larger objects. Hal Levison put up a spirited defence of their result,
but two years after the publication of the original paper another
appeared with the deceptively banal title of An analysis of the statis-
tics of the Hubble Space Telescope Kuiper Belt object search. In this
paper, Michael Brown, Shrinivas Kulkarni and Timothy Liggett con-
sidered how many false objects might appear due to various types of
random noise and stated quite bluntly that according to their calcula-
tions the uncertainties in the number of false objects exceeded by a
large factor the number of objects that Cochran and co-workers
claimed to have detected. They concluded that The detection of comet
sized objects in the Hubble Telescope dataset is therefore not possible.
This was a throwing downof the scientic gauntlet ina quite spec-
tacular manner and it was not long before Cochrans teammade a
riposte witha paper entitled The calibrationof the Hubble Space
Telescope Kuiper Belt object search: setting the record straight. In
this paper they described howthey had implanted articial objects
witha range of brightnesses into their original data and thensearched
for themusing the same methods they had used before. What they
found was that the limiting magnitude of the survey, whichis gener-
ally dened as the level of brightness at whichhalf of the objects
knownto be inthe data are actually found, was a Vmagnitude of 28.6.
As a further check, they brought inanother astronomer to performthe
same analysis. Peter Tamblynused software whichhad beendevel-
oped independently of the original searchteam, but his re-analysis
produced essentially the same result, a limiting magnitude of 28.4.
Although he has not published a rebuttal of Cochrans rebuttal,
Michael Brown remains unconvinced. There is no point in continu-
ing to debate this in the literature, he said, We want people to look at
the papers and make up their own minds as to who is right. This is
good healthy scientic debate and it is how the scientic method
works. The key issue is repeatability, not arguments about who knows
how to do statistics properly. Based on their earlier success, Levison
and colleagues were awarded more time on the Hubble telescope to
repeat the experiment with more images. The second set of observa-
tions were made in the last week of August 1997 using a somewhat dif-
B
e
y
o
n
d

P
l
u
t
o
148
ferent observing strategy. Instead of staring at a single point and
letting the target objects drift across the images, this time they tried
tracking the telescope at the typical rate of motion for a Kuiper Belt
object. This turned out to be rather more complicated than they
expected. As the centre of the eld of view moved slowly across the
sky, it was necessary for the satellite to switch from one guide star to
another as the observations progressed. This made reconstructing the
nal images very difcult and a lot of time had to be spent just under-
standing the details of exactly how the Hubble Telescopes pointing
system worked. This extra work delayed the processing of the data
and by early 2001 no results of the second experiment had been pub-
lished. Time will tell if the original result will be vindicated.
Whatever the eventual fate of the Hubble Telescope result, the data
from the other searches have established a fairly good luminosity
function for trans-Neptunian objects with V magnitudes in the range
from about 21 to 26. Converting this into a distribution of sizes and
masses requires correction for the aforementioned selection effects
and a model of how the real objects are actually distributed in space.
Most estimates are starting to converge on some values which seem
reasonable. There would appear to be about 70 000 objects bigger than
100 km in diameter in the region between 30 and 50 AU, which encom-
passes both the Plutino and Classical Kuiper Belt populations. That
this represents a huge population is illustrated by the fact that there
are only a couple of hundred main belt asteroids in this size range.
For a reasonable size distribution, the population of 1 km sized trans-
Neptunian objects inside about 50 AU is probably about 100 million.
Assuming these objects are indeed icy planetesimals, this corre-
sponds to a total mass equal to about 0.2 of the mass of the Earth, a
value in encouraging agreement with early estimates, such as those
by Whipple, of the likely total mass of the present-day Kuiper Belt.
What of the population of scattered disc objects like 1996 TL
66
? This
is an even trickier problem, since such objects are very difcult to
nd. By mid 2000, just a handful of such objects were known. A
measure of the problem is that one of these objects, 1999 CY
118
, is
bright enough to be detected by a 3.6 m telescope over only 0.24% of its
almost 1000 year orbit. Based on these rather small numbers, and
some careful simulations of how a population of objects might be dis-
covered by the survey teams involved, Chad Trujillo estimates that
there are about 30 000 objects in the scattered disc. This amounts to a
further 0.05 Earth mass of material.
N
u
m
b
e
r
s

a
n
d

s
i
z
e
s
149
1
0
-
5
1
0
-
3
1
0
-
1
1
0
1
1
0
3
1
0
5
1
8
2
0
2
2
2
4
2
6
2
8
J
e
w
i
t
t
,

L
u
u

a
n
d

T
r
u
j
i
l
l
o

(
1
9
9
8
)
J
e
w
i
t
t
,

L
u
u

a
n
d

C
h
e
n

(
1
9
9
6
)
L
u
u

a
n
d

J
e
w
i
t
t

(
1
9
9
8
)
U
p
p
e
r

L
i
m
i
t
s
C
o
c
h
r
a
n

e
t

a
l
.

(
1
9
9
5
)
G
l
a
d
m
a
n

e
t

a
l
.

(
1
9
9
8
)
C u m u l a t i v e S u r f a c e D e n s i t y [ d e g
- 2
]
R
e
d

L
i
m
i
t
i
n
g

M
a
g
n
i
t
u
d
e
L
J
9
8 G
9
8
C
o
c
h
r
a
n

e
t

a
l
.
?
F
i
g
u
r
e

8
.
1
A

c
u
m
u
l
a
t
i
v
e

l
u
m
i
-
n
o
s
i
t
y

p
l
o
t

o
f
t
r
a
n
s
-
N
e
p
t
u
n
i
a
n
o
b
j
e
c
t
s
.
T
h
e

h
o
r
i
z
o
n
t
a
l

a
x
i
s

i
s
r
e
d

m
a
g
n
i
t
u
d
e

w
i
t
h

b
r
i
g
h
t
e
r
,
a
n
d

s
o

o
n

a
v
e
r
a
g
e

l
a
r
g
e
r
,
o
b
j
e
c
t
s
t
o

t
h
e

l
e
f
t
.
T
h
e

v
e
r
t
i
c
a
l

s
c
a
l
e

i
s
t
h
e

n
u
m
b
e
r

o
f
o
b
j
e
c
t
s

f
o
u
n
d

p
e
r
s
q
u
a
r
e

d
e
g
r
e
e

o
f
s
k
y

w
h
i
c
h

a
r
e
b
r
i
g
h
t
e
r

t
h
a
n

t
h
a
t

m
a
g
n
i
t
u
d
e
.
D
a
t
a

f
r
o
m

v
a
r
i
o
u
s

g
r
o
u
p
s

a
r
e
p
l
o
t
t
e
d

a
n
d

t
h
e
y

g
i
v
e

r
o
u
g
h
l
y
t
h
e

s
a
m
e

s
l
o
p
e
.
N
o
t
e

t
h
a
t

t
h
e
p
o
i
n
t

i
n

t
h
e

t
o
p

r
i
g
h
t

c
o
r
n
e
r
,
d
e
d
u
c
e
d

b
y

A
n
i
t
a

C
o
c
h
r
a
n

a
n
d
c
o
-
w
o
r
k
e
r
s

u
s
i
n
g

d
a
t
a

f
r
o
m

t
h
e
H
u
b
b
l
e

S
p
a
c
e

T
e
l
e
s
c
o
p
e
,
i
s
a
b
o
v
e

e
v
e
n

t
h
e

s
t
e
e
p
e
s
t

l
i
n
e
.
(
D
a
v
e

J
e
w
i
t
t
.
)

Things that go
bump in the
dark
Analysis of the many surveys of the trans-Neptunian population
lead to another interesting conclusion; the Kuiper Belt is quite
thick. That is to say that it extends a considerable distance above
and below the invariable plane of the solar system. Exactly how
thick the belt is is unclear since most of the surveys have been concen-
trated in the ecliptic region. Objects in orbits with low inclination, in
other words those that are orbiting in much the same plane as the
rest of the planets, spend all their time near the ecliptic. Even objects
in orbits that are inclined relative to this plane still cross it twice on
each orbit and spend at least some of their time there. Not surpris-
ingly then, few searches have been done at distances far away from
the ecliptic. Thus, the population of objects in highly inclined orbits,
which denes the thickness of the disc, is not well known. Despite
these uncertainties it is safe to say that the Kuiper Belt is at least
30 degrees thick and that it may actually be much thicker than
this.
This thickness has important consequences. Unlike objects all
moving around in near-circular orbits in the same plane, like dancers
in a ballroom, objects whose orbits cross each other at more drastic
angles can collide at quite considerable speeds. Don Davis and Paulo
Farinella have modied computer models which they originally devel-
oped for studies of the main asteroid belt to investigate the role of col-
lisions in the trans-Neptunian region. Surprisingly, the time between
collisions turns out to be roughly the same in both regions; the much
larger trans-Neptunian population balances out the larger volume of
space involved. The most signicant difference is the relative speeds
of the collisions. In the asteroid belt a typical collision occurs between
two rocky objects at about 56 km per second. In the trans-Neptunian
C
H
A
P
T
E
R

9
151
region icy objects strike each other at speeds of only 11.5 km per
second.
Davis and Farinellas computer simulations show that once a popu-
lation of large trans-Neptunian objects had formed, it was essentially
unchanged over the history of the solar system. At the impact speeds
involved even a collision between two equal-sized objects is not violent
enough to disrupt them if they are more than about 100150 km in
diameter. These larger objects will, however, probably be heavily
cratered by numerous smaller impacts. It is also possible that they are
not solid bodies, but comprise rubble piles of very large pieces held
loosely together by gravity. However, at small sizes the story is quite
different. Objects smaller than about 100 km in diameter are broken
up by collisions. The collisions produce a cascade of fragments
moving at speeds that differ in velocity by tens to hundreds of metres
per second compared with the original speed of the parent body. This
could allow the fragments to move away from the parent bodys origi-
nal orbit and might well shift them from a gravitationally stable
region into one from which they could evolve inwards towards the
Sun. If this happened then some of the fragments could eventually
become visible as comets. Davis and Farinella estimate that almost all
the small trans-Neptunian objects which we see today, and which will
eventually enter the inner solar system as comets, are not primordial
at all. They are multigeneration fragments of much larger bodies.
They also calculate that about ten new comet-sized (i.e. 110 km diam-
eter) fragments are being produced every year by collisions in the
Plutino and classical Kuiper Belt region. However, the amount of col-
lisional evolution falls off signicantly as the distance from the Sun
increases. A primordial population of objects at roughly 80 AU (if
such a population exists) will not have undergone signicant colli-
sional changes since it was formed.
Don Davis admits that astronomers do not yet understand the
structure of comet-like bodies well enough to know how they would
behave during collisions at velocities of hundreds of metres per
second. He and other scientists are working on this problem by carry-
ing out laboratory experiments involving high-speed impacts into icy
targets. One of these scientists is Eileen Ryan, from Highlands
University in New Mexico. Her speciality is making, and then break-
ing, snowballs. Ryan says she was fascinated by astronomy from an
early age and was always focused on some kind of a career in science.
She did an undergraduate degree in physics and, although she started
B
e
y
o
n
d

P
l
u
t
o
152
doing postgraduate work in astronomy in New Mexico, it was not long
before she changed course slightly. She moved to the University of
Arizona, and into planetary science. While a graduate student in
Arizona, she found herself working some of the time at the Planetary
Science Institute, a small non-prot research group just off the uni-
versity campus whose staff included Don Davis and several other
well-known asteroid researchers. The institute had just started an
experimental programme to study asteroid collisions and before long
Ryan, who admits to having been at the right place at the right time,
was essentially running this project. She found the work interesting
and exciting and was very surprised to realise how little was actually
known about what happens when two astronomical objects hit each
other. Studies of asteroid collisions led naturally to investigating the
collisional physics of icy objects in order to understand what was hap-
pening in the trans-Neptunian region.
The experimental runs take place at the Ames Research Center at
Moffett Field in California. Here, NASA has built a special test facility,
the Ames Vertical Gun Range, for experimental work on high-speed
impacts. The facility comprises a vacuum chamber, about 2 meters
across, in which the targets are placed, and a variety of guns. There
are several different kinds of guns which allow projectiles to be red
over a wide range of velocities depending on what the experimenter
requests. Targets are installed in the chamber and then bombarded by
high-speed projectiles from different angles. Ultrafast cameras, taking
T
h
i
n
g
s

t
h
a
t

g
o

b
u
m
p

i
n

t
h
e

d
a
r
k
153
Figure 9.1 Eileen Ryan, whose
impact experiments at the NASA
vertical gun range are used to help
understand how objects in the
trans-Neptunian region evolve.
(Eileen Ryan.)
400 frames per second and sometimes supplemented by a video
camera storing 1000 frames per second, record each test. The Ames
gun has been used to study cratering on the Moon and planets by
ring projectiles at sand or rock targets, and for studies of collisions
between asteroids.
For their tests Ryan and Davis needed targets about 10 cm across,
which had to be representative of trans-Neptunian objects.
Unfortunately, no-one truly understands what the real things look like
or how they are made up. Ryan thinks it is likely that they are porous,
uffy or grainy in structure, but it is not known if all the constituent
pieces are of the same size, or if they cover a range of sizes.
Accordingly, the experiments were done using a variety of targets
which were made using ice blocks bought from a local supermarket.
For targets comprising an aggregation of particles with a wide range
of sizes, Ryan took some ice chips and hammered them into frag-
ments. She then squeezed the fragments together as if making a snow-
ball. Other targets were composed only of large pieces, each piece
being an ellipsoid about 2.5 cm long (basically ordinary ice cubes).
Others were put together with 0.5 cm grains. To make very-small-
grained targets, they put ice in a kitchen blender and ground it up
before compacting the resulting icy powder. To nish the job, each of
the targets was allowed to melt slightly and then pushed into a mould
and refrozen. Ryan admits that, Making the targets doesnt sound
very scientic, but in real life your clever ideas actually have to be
implemented somehow.
A variety of projectile types, including pellets of aluminium, solid
ice and fractured ice, are red into the simulated Kuiper Belt objects.
The aluminium pellets are used since similar projectiles had been
used in the past for experiments with rock targets and using the same
type of projectile simplied comparisons between the different exper-
iments. Icy projectiles were more realistic, but rather harder to work
with, since the golf-ball-sized ice projectiles tended to break during
the launching process. If this happened the projectile would arrive as
a stream of ice particles rather than a single object.
Each experimental run lasts about two weeks and comprises thirty
or forty shots in total. For each shot Ryan and her group begin by
lining the chamber with padding to stop extra unwanted fragmenta-
tion caused by pieces ying off the target and then hitting the
chamber walls (although experience shows that most fragments do
not hit the walls, but just fall to the ground). The chamber oor is
B
e
y
o
n
d

P
l
u
t
o
154
covered with a plastic tarpaulin. After the shot, as soon as the air has
been pumped back into the chamber, the team rushes in. They scoop
up the fragments and then run upstairs to a cold room where the frag-
ments are preserved. They try to collect every fragment weighing
more than half a gram, and they usually manage to gather up more
than 80% of the original mass of ice. The smaller fragments melt
almost immediately and are lost. During the day, or the next day, the
fragments in the cold room are measured, weighed and photographed.
This allows the fragment sizes and masses, and the mass of the largest
fragment to be recorded. The fragment velocity distribution is
deduced from examining the lm and video recordings taken during
each experiment. Ryan takes the results of each shot and uses a two-
dimensional hydrodynamic computer code to model the result of the
collision and then she compares the outcome calculated by the com-
puter with what actually happens in the test chamber. If the results of
the computer model match the data, it is then possible to extrapolate
upward by seven, eight or even nine orders of magnitude to estimate
the effects of collisions between larger bodies.
According to Ryan, this sort of information is critical to nd out
how collisions have affected the Kuiper Belt and what is happening to
objects in those regions. Astronomers need to know what kind of frag-
ments are produced, what the mass distribution is, at what speeds the
T
h
i
n
g
s

t
h
a
t

g
o

b
u
m
p

i
n

t
h
e

d
a
r
k
155
Figure 9.2 A test target representing a planetesimal in the trans-Neptunian region. It
was assembled from a bag of commercially available ice cubes. (Eileen Ryan.)
fragments are ejected and what they hit next. The results are interest-
ing, and not very intuitive. It appears that porous ice targets behave as
strongly as solid ice during collisions, even though the porous ice
targets are weak when subjected to more gradual forces. The reason
seems to be that the sudden pulse of energy from the collision is dissi-
pated by the empty spaces within the porous structure. This delays
and confuses the shockwave as it travels through the material after
the impact. It is a process that is very complicated to model, but it is
essential that we understand it if we are ever to understand the colli-
sional evolution of the objects in the outer solar system.
Eileen Ryans experiments mostly concern collisions in which the
target is shattered into many fragments, but smaller collisions must
be important too. Given all the other uncertainties, it is not possible to
be sure exactly how frequent these collisions are, but it seems likely
that a large object, say one about 100 km in diameter, is struck by a
1 km object every 300 years or so. When a smallish object crashes into
a much larger one then it is presumably destroyed completely and
forms a crater. The impact will excavate material from the target,
some of which will land again and form a blanket around the crater.
Similar effects are seen in the form of the bright rays which extend
outwards from a number of recent craters on the Moon. This impact
gardening of the surface (also sometimes called space weathering)
B
e
y
o
n
d

P
l
u
t
o
156
Figure 9.3 The fragments of one of Ryans test targets after an impact experiment. These
results suggest that small trans-Neptunian objects may well be irregular. (Eileen Ryan.)
may be responsible for some of the colour diversity reported by
various observers making photometric observations. However, since
even the largest objects in the trans-Neptunian region are quite small
by planetary standards, they do not have particularly strong gravita-
tional elds. Much of the ejected material is probably thrown off into
space as dust and ne debris.
This dust will not remain in the trans-Neptunian region forever, or
even for very long. A variety of physical effects act on very small par-
ticles and cause them to be either blown out of the solar system or to
spiral inwards towards the Sun. One of these is called
PoyntingRobertson drag. This occurs because sunlight striking a
dust grain in solar orbit appears (to the dust grain) to be arriving from
slightly in front of the particle. The arrival of this stream of photons
on its front face provides a braking force which slows the grain and
causes it to spiral into the Sun. Dust is also removed by impacts with
other grains already orbiting in the locality or with interstellar parti-
cles passing through the solar system. All of these effects conspire to
remove dust from the trans-Neptunian region on timescales of
around a million years. So while there is probably a cloud of dust in
the Kuiper Belt, it is not likely to be very dense.
That any dust in the trans-Neptunian region must be quite thinly
spread out is conrmed by the fact that no-one has yet been able to
observe it. Fred Whipple rejected searching for sunlight reected
from the dust since it would be too faint to see against the other
sources of faint glows in the night sky. More recent attempts have
tried to use not light emitted from the Sun and then scattered back
towards the Earth, but rather the thermal infrared radiation emitted
by the particles themselves. This experiment would be impossible to
do from the ground, but in 1990 the American Cosmic Background
Explorer (COBE) spacecraft mapped the entire sky in the far infrared
to search for cosmological signatures of the Big Bang. Along the way,
COBE also mapped out the zodiacal background and numerous other
faint sources of infrared radiation. These sources of far infrared radi-
ation had to be modelled so they could be removed from the data
before it was searched for the cosmological background which was the
main objective of the mission. COBE scientists did not nd any evi-
dence for a band of trans-Neptunian dust. While the limit which this
failure placed on the amount of dust that might be present was not
very strict, it was at least consistent with estimates based on the rate
at which dust is being produced by collisions.
T
h
i
n
g
s

t
h
a
t

g
o

b
u
m
p

i
n

t
h
e

d
a
r
k
157
To date, the only observational measurements of dust in the Kuiper
Belt have come from an unusual source, spacecraft which are speed-
ing away from the Sun after their spectacular encounters with the
giant planets in the late 1970s and 1980s. Pioneer 10 ew past Jupiter
in 1973 and has since penetrated more than 10 AU into the trans-
Neptunian region. The spacecrafts dust-detection system, designed to
measure dust in the asteroid belt, is no longer operating, but its
hydrazine fuel tank has been used to place a limit on the number of
medium-sized particles in the region through which it has been travel-
ling. Scientists have calculated that it would take a particle a few mil-
limetres across to puncture the 42 cm diameter tank and the fact that
the tank has survived so far limits the number of such particles.
Although the limit which this result provides is not very strict, being
about one tenth of an Earth mass, the result is an interesting one
since the size range in question is difcult to measure by any other
method. Particles in this size range are too large and too far apart to
produce detectable diffuse emission and much too small to detect
directly. The experimenters have developed their idea to suggest that
other parts of the spacecraft, such as portions of the communications
system, wiring bundles and some of the other electronics, might also
be tested to see if they have suffered impact damage. Unfortunately,
the Pioneer spacecraft is now so short of electrical power that there is
none to spare. It is no longer possible to turn on selected instruments
for even a short time just to see if they still work.
The two larger and more complicated Voyager spacecraft do not
have dust experiments, but they have detected trans-Neptunian dust.
Both Voyagers are equipped with plasma detectors designed to
measure charged particles trapped by the magnetic eld of the giant
planets. These detectors also register some plasma when a high-speed
particle strikes the spacecraft and is vapourised on impact. So, by
using the Voyager itself as a detector, and recording the bursts of
plasma produced by the impacts, a crude dust detection experiment
has been improvised. Of course, it is far from a perfect system. In par-
ticular, the range of particles it detects is rather limited; small grains
do not produce measurable effects and large ones are very rare. Still,
until other spacecraft arrive in this region, it is the best that can be
done.
The results of the Voyager experiments are intriguing. One space-
craft is dropping below the ecliptic plane and the other is climbing
above it. Both spacecraft did detect dust impacts as they started to tra-
B
e
y
o
n
d

P
l
u
t
o
158
verse the trans-Neptunian region, but in both cases these detections
have now stopped. Voyager 2 data cut off at 33 AU and Voyager 1 made
its last detection at 51 AU. It may be that the instruments have just
ceased working, not unreasonable for equipment that has been in
space for close to 25 years, but perhaps the dust has really thinned out
to the point that it is not being detected.
Dynamical arguments, sky surveys and spacecraft dust experi-
ments all suggest that the present-day Kuiper Belt contains less than
one Earth mass of material. However, there are strong indications
that this has not always been the case. The present mass of material
may be only a few per cent of the original solar nebula at this distance
from the Sun. There are several lines of evidence that this must have
been so. The rst harks back to the original papers of people like
Edgeworth and Kuiper and concerns the overall density of the solar
system at increasing distances from the Sun. If it were possible to
grind up all of the non-volatile material in the giant planets and
spread it out in rings at each planets heliocentric distance, then there
would be a fairly gradual decline in mass density with increasing dis-
tance from the Sun until the orbit of Neptune was reached. After
Neptune there would be a very sudden and dramatic drop. It was this
sharp edge to the solar system that made Kuiper and Edgeworth spec-
ulate about a trans-Neptunian disc. Even though such a disc has now
been found, the fundamental problem remains. If all the objects and
dust in the disc are added together there still doesnt seem to be any-
thing like as much material as there should be.
Another line of evidence comes fromthe very existence of large
Kuiper Belt objects and this has beeninvestigated by the American
AlanStern. AlanSternis a true aerospace enthusiast. He has degrees
inphysics and aerospace engineering and is a qualied pilot of both
powered aeroplanes and gliders. He graduated witha PhDinplanetary
astronomy fromthe University of Colorado and has since been
involved innumerous airborne and space astronomy projects. Today
he is the director of Space Studies at the SouthWest ResearchInstitute
inBoulder, Colorado. Sternhas long beeninterested inthe outer solar
system. Using computer simulations he has shownthat the rate of col-
lisions ina disc withthe same mass as the present-day Kuiper Belt is
far too slowto growobjects the size of 1992 QB
1
. The density of mater-
ial this far fromthe Sunis so lowthat there is simply not enoughtime
since the solar systemformed for comet-sized planetesimals to accrete
into objects several hundred kilometers indiameter. To reconcile this
T
h
i
n
g
s

t
h
a
t

g
o

b
u
m
p

i
n

t
h
e

d
a
r
k
159
B
e
y
o
n
d

P
l
u
t
o
160
0
0
20
20 10
10
10
10
10
10
10
10
40
40
60
60
30
30
50
50
70
70
80
80
DISTANCE (AU)
DISTANCE (AU)
L
O
G
(
g
m
/
c
m


)

COLLISIONALLY
EVOLVED ZONE
DYNAMICALLY
AND
COLLISIONALLY
DEPLETED ZONE
N U
U
? ?
?
?
?
t
t
o
now
S
S
J
J N
1
0
-1
-2
-3
-4
-5
2
Figure 9.4 The mass density across the outer solar system which would result if the
giant planets were ground back up into dust and spread out as rings at their present dis-
tance from the Sun. The line labelled t
o
shows the estimated density of the solar nebula
early in its history. The solid line t
now
represents the situation today. Even taking into
account the Kuiper Belt objects between 30 and 50 AU, there seems to be a sudden drop in
density after the planet Neptune. Alan Stern and others have shown that this region is
depleted by erosive collisions, but speculate that beyond 50 AU the density of material
may begin to rise again. (Alan Stern/Astronomy Society of the Pacic.)
problemwiththe fact that suchobjects clearly do exist, Sternsuggests
that at some time early inits history the Kuiper Belt must have been
more massive. He estimates that the regionbetween30 and 50 AUmust
once have contained at least ten, and perhaps as muchas fty Earth
masses of material. Interestingly, this is about the amount of material
youwould get by extrapolating the mass density of the giant planets
out into the trans-Neptunianregion.
Stern also concluded that even if such a massive disc did exist early
on in the solar systems history, the growth of large objects could only
have occurred if the objects in the disc were in orbits of fairly low
eccentricity and inclination. Only if this was true would the mutual
encounters between planetesimals be gentle enough that there would
be a good chance that the objects would stick together. The present-day
trans-Neptunian disc is quite thick and, as shown by Don Davis and
Paulo Farinella, collisions within the disc are more likely to lead to
erosion than to growth. So the objects we see there today must have
been formed quite early on, before the growth of Neptune and its sub-
sequent gravitational inuence on the forming disc could be
signicant. The existence of numerous Kuiper Belt objects with diam-
eters of a few hundred kilometers suggests a sort of race took place
between the planetesimals in the trans-Neptunian region and the
proto-Neptune. For perhaps several hundred million years, the plan-
etesimals were trying to grow as quickly as possible before Neptune
increased too much in mass. Once Neptune grew to the point that its
gravity stirred up the orbits of the nearby planetesimals, their mutual
collisions became destructive and any further growth stopped.
Finally, there is the existence of Pluto and its moon Charon. How
did this unlikely pair ever form and come into orbit around each
other? Pluto and Charon are both small icy worlds. They presumably
formed in the outer solar system by the gradual accretion of smaller
planetesimals, just like the other objects in the trans-Neptunian
region. The problem is that there is no known process which would
allow Charon to form in orbit around Pluto. Instead, it appears most
likely that two quite large bodies formed independently and then col-
lided. The result of such a collision could have thrown huge amounts
of material into orbit around what was left of Pluto and this material
eventually re-accreted to form Charon.

For a variety of reasons this

T
h
i
n
g
s

t
h
a
t

g
o

b
u
m
p

i
n

t
h
e

d
a
r
k
161

The Earths moon is believed to have been formed in a similar way when an object the
size of Mars struck the proto-Earth and blasted material into Earth orbit.
is an attractive model but for one thing, the likelihood of such a colli-
sion is very, very small. The probability is less than one in a million
over the age of the solar system. Now, million to one chances do come
off, occasionally, but using such rare events to explain things tends to
make astronomers nervous.
One solution to the mysterious origin of the PlutoCharon binary
is to assume that there were a lot more Pluto-sized planetesimals early
in the history of the solar system. By increasing the number of
objects, the odds of a collision are lowered to something a bit more
reasonable. Alan Stern argued, before a single Kuiper Belt object had
been found, that an ancient population of thousands of 1000 km sized
ice dwarfs could be invoked to explain several puzzling features of
the outer solar system. The unusual tilt of the planet Uranus, which
lies more-or-less on its side as it orbits the Sun, could be explained by
an impact by one or more large bodies late in the planet-building
process. Neptune also has quite a large axial tilt, about 29 degrees,
which can be explained in the same way. Neptune has a large satellite,
Triton, which orbits the planet in the wrong direction compared with
the sense of rotation of the planet itself. Triton could be an ice dwarf
that wandered close to Neptune. However, to be captured into orbit
around Neptune, Triton must lose energy. Possible ways this might
have happened include gas drag in the forming planets outer atmos-
phere or a collision with a small primordial satellite of the planet.
Either way, Triton has to make a very close approach to Neptune
which is, on the face of it, quite improbable. Every one of the scenar-
ios outlined here are possible in isolation, but the chance that they
would all occur is very low. The fact that they did all occur suggests
that there was once a much larger population of these icy dwarfs. The
existence of this large population greatly increased the probability of
what are otherwise very unlikely events.
What was the fate of the remaining ice dwarfs? Like the many
smaller objects formed in the outer planet region, they were probably
ejected by gravitational interactions with the forming giant planets.
Some may have escaped the solar system entirely, the remainder were
either incorporated into planets or were sent into the Oort Cloud and
the scattered disc. There is no difculty in doing this; even though the
objects are orders of magnitude larger than typical comet nuclei they
are still ten thousand times less massive than either Neptune or
Uranus and so easy to banish into the outer darkness. The existence of
such objects in the trans-Neptunian region is not ruled out by the
B
e
y
o
n
d

P
l
u
t
o
162
present data. The cumulative luminosity functions worked out to date
do not rule out the existence of between several and a few dozen Pluto-
sized objects still waiting to be discovered.
All this points to a nal question about the Kuiper disc that still
needs to be resolved. What is the density of material beyond about
50 AU, the limit of most ground-based surveys to date? We have seen
that the density of material in the 3050 AU region is actually a lot
lower than would be expected based on an extrapolation of the density
T
h
i
n
g
s

t
h
a
t

g
o

b
u
m
p

i
n

t
h
e

d
a
r
k
163
Figure 9.5 Neptunes large satellite Triton. This icy world may be one of the last sur-
vivors of a once more numerous population of ice dwarf mini-planets in the outer solar
system. (NASA.)
in the region of the giant planets. This depletion can be explained by
ejection of material during the planet-building process and the steady
removal of mass ever since as erosive collisions grind down the trans-
Neptunian objects. However, beyond about 50 AU the gravitational
inuence of Neptune is negligible and any planetesimals formed in
these distant regions will not have been stirred up by the planets.
These distant objects would probably remain in circular, low-inclina-
tion orbits where high-speed, erosive collisions are rare and they
should still exist today.
Although the mass density beyond 50 AU must have been lower
than it was closer to the Sun, if the 3050 AU region once contained
more than ten Earth masses of material, then the outer regions of the
protoplanetary disc must also have been quite massive. Simulations of
the growth of objects in this region by Stern and his colleague Joshua
Colwell have shown that it is possible to grow objects up to several
hundred kilometres in diameter in less than the age of the solar
system. Although the growth proceeds more slowly so far from the
Sun, the lack of any excitation of the orbits by Neptune means that
accretion can continue for much longer. Indeed, it may still be going
on today. Stern and Colwell argue that since the conditions in the
region between 50 and 100 AU from the Sun are not erosive, then the
present-day density of material may begin to rise again, quite rapidly,
this far from the Sun. If this is indeed the case, then the Kuiper Belt as
we know it today may in fact be a local dip in density. There may be a
wall of material yet to be found still further out.
To date, no sign of such a wall has been found and increasing
numbers of people are beginning to suspect that it never will be.
Simulations of the rate at which objects are being discovered show
that by now at least a few objects in circular orbits beyond 50 AU
should have been found. None have yet turned up. It may be that these
more distant objects are on average smaller or darker, and so doubly
harder to nd, than those closer in. However, according to Chad
Trujillo, to explain his deep survey data in this way would require
some ridiculous assumptions. The sudden drop in discoveries at 50 AU
would require that the more distant objects were all ve times darker,
or somehow systematically much smaller, than the objects just nearby.
R. L. Allen, together with Gary Bernstein, is in broad agreement with
Trujillo. Their data suggest that the Classical Kuiper Belt does not
extend beyond about 55 AU. In contrast, Brett Gladman feels that since
distant objects are so much harder to nd he cant rule out a gradual
B
e
y
o
n
d

P
l
u
t
o
164
fall off in objects with distance. It may also be that the protoplanetary
disc did not extend outwards to great distances after all. Perhaps the
disc was truncated by some process. Stars form in clusters and a close
encounter between the Sun and one of its siblings might have
removed the outer regions of the protoplanetary disc. Luckily, this is
not an unanswerable question. Encounters between the Sun and a
passing star can be modelled on computers and used to predict details
of how the ensemble of objects beyond about 50 AU should appear. In
particular, Shigeru Ida, from Japans Tokyo Institute of Technology,
and his collaborators have shown that an encounter with another star
would tend to increase both the inclination and eccentricity of the
orbits of distant objects. If such distant trans-Neptunian objects exist
then once a reasonable number of their orbits have been determined,
this hypothesis can be tested.
T
h
i
n
g
s

t
h
a
t

g
o

b
u
m
p

i
n

t
h
e

d
a
r
k
165
Dust and discs
The discovery of the Kuiper Belt has established that the planetary
region does not stop at Pluto, but that it extends far further into space
than originally thought. How does the Suns trans-Neptunian disc
compare with the structures astronomers see around other nearby
stars?
Until quite recently astronomers thought that most stars had long
since blown away any remnants of the dust cloud from which they
were formed and that they now existed in more-or-less splendid isola-
tion. Mature stars lie on what is called the main sequence, a stable
period in the life of a star that corresponds roughly to adulthood in
humans. However, hints that main sequence stars might not be that
simple came during the rst few weeks of the mission of the Infrared
Astronomical Satellite (IRAS) in 1983. IRAS was an international
project to survey the sky in a number of far-infrared wavelength
bands that cannot be studied from the ground because of absorption
in the Earths atmosphere. The satellite was launched in January 1983
and, as with any new astronomical instrument, one of the rst tasks
facing the IRAS team was to check the calibration of the satellites
detectors. The way chosen to do this was to observe a number of stars
whose properties were thought to be well understood and whose far-
infrared uxes could be estimated from ground-based observations at
other, shorter wavelengths.
One of the stars chosen was Vega, or Alpha Lyra, the brightest star
in the constellation of the Lyre. Vega is a well-observed main
sequence star which lies quite close to the Sun. It was believed to be of
spectral class A0, just an ordinary hot star which ought to behave like
a black body with a temperature of about 9850 K. Since the ux from a
black body of any given temperature is quite easy to calculate, stars
C
H
A
P
T
E
R

1
0
167
like Vega are often used as calibration sources for astronomical
instruments. Vega in particular is used to dene the zero point of the
astronomical magnitude scale. All astronomical magnitudes in the
ultraviolet, visible and infrared regions are linked to the brightness of
Vega one way or another. Accordingly, Vega was a prime target early
in the IRAS mission. However, quite soon after launch, it appeared
that something was wrong; observations of Vega showed that the
IRAS detectors seemed to be over-responsive to long wavelengths.
This was a puzzle, and one which demanded the immediate atten-
tion of the scientists and engineers at the IRAS ground station at the
Rutherford and Appleton Laboratory in Didcot, England. Luckily, it
did not take long to nd out what was going on. Observations of a
number of other sources came out more or less as expected and it was
soon clear that the problem was not with IRAS, but with Vega. The
star is not a simple black body. It has excess emission in the far
infrared which had escaped detection from the ground. Excess
infrared radiation is quite commonly associated with young (i.e. pre-
main sequence) stars. It is caused by hot dust close to the star which is
heated up by starlight and then radiates heat away in the infrared.
However, Vega, a well-evolved main sequence star, should have lost all
of its original protostellar dust long ago. What was more, the IRAS
data indicated that the dust was cool, only about 50 K or so. This low
temperature implied that the dust was not close to the star, but that it
was in a disc some distance away from Vega itself.
Once the big hint had been dropped by the discovery of the Vega
disc, it did not take long before infrared excesses were found around
other nearby main sequence stars. Three stars stood out almost at
once, Beta Pictoris, Alpha Pisces (Fomalhaut) and Epsilon Eridani.
Eventually, a closer examination of IRAS database showed that
numerous otherwise apparently ordinary stars also had infrared
excesses, all suggestive of the presence of cool dust. What was more,
in almost every case, the age of these stars greatly exceeded the theo-
retical lifetime of dust in the system. If the dust was just a remnant of
the original protostellar cloud from which each star had formed, it
should have long since been removed by the same effects that are
presently clearing out dust from the Suns Kuiper Belt. The implica-
tion was clear, something must be replenishing the dust around these
stars. Could it be mutual collisions within a population of comets and
asteroids?
The detectors on the IRAS satellite had very low spatial resolution;
B
e
y
o
n
d

P
l
u
t
o
168
at best it was only a few arcminutes. This was perfectly adequate for
a sky survey mission, but was very poor by the standards of a normal
ground-based telescope. So, although IRAS could detect the presence
of dust around Vega and other stars, it could reveal very little about
how the dust was actually distributed. Just about the only conclusion
that seemed to be safe was that the temperature of the dust suggested
that it was in a disc with a hole in the middle. The next development
was not long coming. Brad Smith at the University of Arizona and
Rich Terrile at the Jet Propulsion Laboratory in Pasadena, California,
began a programme to look at what were already being called Vega
Excess Stars using a camera tted with an optical coronagraph. A
coronagraph is a device used to block out the light from a bright astro-
nomical source so that any fainter objects which might be nearby can
be seen. The process is very similar to a car driver blocking the head-
lights of an approaching car with a hand in order to see the much
fainter light from the reectors which mark out the roadway.
Coronagraphs get their name from devices used to block the light
from the Suns disc, so that its outer region, the corona, can be studied,
but they work just as well during regular astronomical observations
looking for faint companions to bright stars.
Smith and Terrile turned a coronagraphic camera towards the star
Beta Pictoris (usually abbreviated to Beta Pic) and hit the jackpot
almost at once. Clearly visible on either side of the star were wings of
reected light. Beta Pic had a disc of dust and gas almost exactly edge
on to us. Particles in the disc were scattering the visible light from the
star back towards the Earth. The disc was huge, spanning 1500 AU
from edge to edge. Closer examination of the images showed that the
disc was not symmetrical and that it was quite thick. The thickness
hinted that there might be planetary-sized bodies embedded in the
disc which were stirring up the material, stopping the disc attening
out.
Since Smith and Terriles discovery many more observations have
been made of the Beta Pic disc. During 1993 Paul Kalas and Dave
Jewitt investigated the asymmetries in the disc in detail and showed
that the dust orbiting between 150 and 800 AU from the star was indeed
asymmetrical. They concluded that the large scale of this asymmetry
meant that it was unlikely to have been caused by planets orbiting
close to the star. They suggested instead that the disc may have been
disturbed by the gravitational eld of a star passing close to Beta Pic
sometime in the last few thousand years. More recent observations
D
u
s
t

a
n
d

d
i
s
c
s
169
have veried the large-scale asymmetry of the disc conrming that it
is brighter on one side than on the other. Observations from the
Hubble Space Telescope also revealed the presence of warps or kinks
in the inner regions of the disc. These inner structures are on scales
similar to the dimensions of our own solar system. They point even
more clearly to the existence of planetary bodies orbiting somewhere
inside the system and gravitationally modifying the structure of the
disc.
Optical images of the other Vega excess stars did not detect any
more discs, but this was not entirely surprising. For a disc to be seen
in normal reected light, it must lie almost edge on to the line of sight.
Such a situation is likely to be quite rare. Circular discs seen close to
face on are very difcult to detect directly, since there is not enough
material along each sightline to reect much light back towards the
Earth. The best way to detect nearly face on discs is not to rely on
reected starlight, but rather to observe at wavelengths where the
warm dust is actually emitting radiation. The dust disc of a typical
Vega excess star has a temperature of about 4070 K. The peak of the
thermal emission from such dust is in the far infrared, but there is
B
e
y
o
n
d

P
l
u
t
o
170
Figure 10.1 Space telescope images of the Beta Pictoris Disc at two different scales. The
top image shows the large-scale structure of the outer disc, the bottom one shows warps in
the very central regions which may result from the gravitational effects of planet-like
bodies close to the star. (NASA/STScI.)
still signicant ux in the somewhat longer sub-millimetre region.
By good fortune, this area of astronomy was just being opened up to
ground-based astronomers.
Millimetre and sub-millimetre telescopes are large dishes rather
like radio telescopes in appearance. The main difference is that since
the sub-millimetre band encompasses much shorter wavelengths than
the traditional radio bands, sub-millimetre telescopes have surfaces
that are much smoother than their radio cousins. Sub-millimetre radi-
ation is strongly absorbed by water vapour and so astronomers inter-
ested in this wavelength region have sought high and dry locations for
their instruments. One such sub-millimetre telescope is the James
Clerk Maxwell Telescope (JCMT) on Mauna Kea. The JCMT has a dish
15 metres in diameter comprising of 276 reective panels. All of the
panels are aligned and supported so that the dish is very smooth; devi-
ations across the surface are no greater than the thickness of a page
from a typical telephone directory. The whole telescope is inside a
weather-tight enclosure which protects it from the elements and main-
tains the critically important shape of the dish.
Sub-millimetre instruments, usually called bolometers although
their function is to do photometry, were soon trained on the Vega
excess stars in an attempt to map out the suspected dust shells at long
wavelengths.

Observations were made from the JCMT and from other

similar telescopes such as the Institut de Radioastronomie
Millimtrique (IRAM) in France and the Swedish-ESO Sub-
millimetre Telescope (SEST) in Chile. Unfortunately, the rst results
were not very encouraging. Sub-millimetre observing was in its
infancy and like early optical and infrared photometers, each instru-
ment had just a single detector viewing the sky. This made mapping
extended emission a very laborious process. To make a map it was nec-
essary to rst point the telescope at the star, then move off slightly to
one side before making a measurement. Once this was done the tele-
scope was pointed to a position on the other side of the star and the
process was repeated. Given enough time, and enough good weather,
a small grid of positions could be mapped around the star. A crude
image could then be reconstructed by computer processing of all the
data. It was a slow and tedious procedure. Mapping faint sources was
limited by both the relatively low sensitivity of the early instruments
D
u
s
t

a
n
d

d
i
s
c
s
171

Most textbooks dene a bolometer as a device which absorbs radiation of all wave-
lengths. By placing lters in front of it it can be used as a photometer.
and by the large beams, and so poor spatial resolution, produced by
even the best sub-millimetre telescopes. Worse still, several groups
trying to make similar observations did not get similar results and
there was some nger pointing as each group tried to defend its own
measurements at the expense of those from its rivals.
This situation was only resolved (if the reader will pardon the
expression) by the development of new sub-millimetre instrumenta-
tion with multiple detector elements of improved sensitivity.
Probably the best known of these instruments is SCUBA, the Sub-mil-
limetre Common User Bolometer Array, installed at the JCMT in
Hawaii. SCUBA has over 100 bolometer detectors, each of which is
about ten times more sensitive than those in the previous generation
of instruments. Since the signal-to-noise ratio of an observation
increases as the square of the sensitivity, and in direct proportion to
the number of detectors, SCUBA is about 10 000 times more efcient
as a mapping instrument than any of the single-element bolometers
which preceded it. SCUBA was an obvious tool with which to attack
the problem of the Vega excess stars and it was not long before several
groups attempted to do just that.
The rst and most spectacular success fell to the wife and husband
team of Jane Greaves and Wayne Holland, staff members of the Joint
Astronomy Centre in Hawaii. They used SCUBA to map the sub-mil-
limetre emission from a number of the brightest Vega excess stars
and soon came up with a very interesting conclusion. Their images
showed that not only did these stars have the dusty discs predicted by
the far-infrared observations from IRAS, but the discs were asymmet-
ric. They had blobs of emission around the circumferences as well as
clearly dened holes in the middle. SCUBA was a new instrument and
one which had been quite difcult to bring into regular use. Since the
presence of asymmetries in the discs was unexpected, considerable
effort was put into checking that the features in the SCUBA images
were real and not an artifact of either the new instrument or its com-
plicated data reduction software. After careful checking, Greaves and
collaborators were convinced; the structures were real.
The SCUBA images immediately explained why the rst sub-mil-
limetre observations had been so difcult to reconcile. Different tele-
scopes, with their inevitably different beam sizes, had seen different
parts of these structured discs. Each had detected different amounts
of radiation. As Jane Greaves put it, If only they had looked further
out they would have seen lots of ux from the outer regions of the
B
e
y
o
n
d

P
l
u
t
o
172
discs, but of course they didnt know that there was anything out
there. They were basically staring through holes in the central
regions of the discs and missing most of the ux. The SCUBA images,
coming as they did about the same time that great strides were being
made in understanding the outer regions of the solar system, sud-
denly made perfect sense. SCUBA, and the other sub-millimetre
cameras coming into service, were staring at systems which had
apparently already started to form planets. Each star was surrounded
by its own population of dust and planetesimals. Even the scales of
the discs were about the same as the Suns Kuiper Belt.
The most spectacular extrasolar dust disc is that around the nearby
star Epsilon Eridani. This star is about 0.8 times the mass of the Sun
and is estimated to be about as old as our own Sun was during the nal
stages of planet formation. Epsilon Eridani seems to have a dust ring
extending between about 35 and 75 AU and containing at least 1% of
an Earth mass of material. This is actually a rather small amount of
dust, compared with the Suns Kuiper Belt, especially for a disc which
is comparatively new and has not yet lost much of its material. Jane
Greaves says this is not an insuperable problem. SCUBA is sensitive
to quite small grains, she says. SCUBA sees mostly particles about
100 microns, or one tenth of a millimetre, in diameter. If the disc
around Epsilon Eridani has evolved to the point that it contains
signicant numbers of centimetre or large sized particles, there could
easily be a lot more mass than we think. The problem with larger par-
ticles, at least as far as SCUBA is concerned, is that they have a lot of
mass, but only a relatively small surface area. Large particles like
these do not radiate all that much energy in the sub-millimetre and
this makes it hard for SCUBA to see them.
What of the structure seen in the Epsilon Eridani dust disc? The
cavity can be explained by the presence of planets whose gravity has
swept up the residual dust and cleared out the inner disc. Our own
solar system is almost devoid of dust inside the orbit of Neptune for
much the same reason. The outer region of the Epsilon Eridani ring
comprises a structure like the Suns Kuiper Belt where mutual colli-
sions between small planetesimals are replenishing the tiny dust
grains and making the disc detectable with SCUBA. As for the asym-
metries in the disc, these may be huge clumps of material which are
being concentrated by one or more forming planets whose gravity is
still sweeping up the orbiting dust. Further evidence that Epsilon
Eridani may have a planetary system was announced in the middle
D
u
s
t

a
n
d

d
i
s
c
s
173
of 2000. While Anita Cochran had been searching for small trans-
Neptunian objects around our own Sun, her husband Bill had contin-
ued with a long-running project to search for large extra-solar planets.
Cochran, and other competing groups studying other stars, did this by
looking for tiny back-and-fro motions caused by the gravitational tugs
from unseen planets around nearby stars. After some twenty years of
monitoring Epsilon Eridani, Cochran believes there is strong evi-
dence that at least one large planet orbits the star. The planet is about
half the mass of Jupiter and has an eccentric orbit which takes it
around the star every 6.9 years. Cochrans discovery adds to the belief
that Epsilon Eridani today may be a snapshot of what our own solar
system looked like a few billion years ago.
Although Epsilon Eridani is the most spectacular of the dust rings
so far detected by SCUBA, it is not the only one. Observations of Vega
B
e
y
o
n
d

P
l
u
t
o
174
Figure 10.2 A SCUBA image of the dust ring around Epsilon Eridani. The ring appears
to have a hole in the centre and clumps of material around its circumference. The extent of
the ring is similar to the Suns Kuiper Belt. (Jane Greaves.)
have also detected the disc around this star. The Vega disc is asymmet-
ric, implying the presence of planets, or at least the progenitors of
planets, around this star as well. The SCUBA images also solved one of
the mysteries surrounding the early sub-millimetre observations of
Vega. Maps made with single element bolometers had apparently cor-
rectly detected that one side of the suspected disc was brighter than
the other, but the data were so uncertain that it was hard to be sure.
Images from SCUBA conrmed the presence of signicant disc asym-
metries which were consistent with the much less reliable, but ulti-
mately correct, measurements made almost a decade earlier.
About the same time as the publication of the SCUBA images, two
American teams announced direct evidence for a dust disc around
another star. Both of these groups were using cameras operating in
the mid-infrared region, that is to say wavelengths of around 10 and 20
microns. As with the sub-millimetre band, it is only recently that mid-
infrared instruments have progressed from single-element detectors
to cameras containing detector arrays with many pixels. In March
1998, two such instruments were turned on the star HR4796A. These
cameras, called OSCIR and MIRLIN, were tted with a 128 pixel
square detector array developed by the Boeing company. MIRLIN was
mounted on one of the Keck telescopes atop Mauna Kea and OSCIR
was tted to the 4 m Blanco telescope at the Cerro Tololo Inter-
American Observatory in Chile.
HR4796A is an A0 star in the southern constellation of Centaurus.
With a V magnitude of 5.78 it is just visible to the naked eye. The star
is believed to be about 10 million years old, quite young compared
with the Sun, but about the right age to be in the late stages of planet
formation. It lies 67 parsecs (about 210 light years) from the Sun. This
is about three and half times further away than Beta Pic and the extra
distance makes detailed observations of it that much more difcult.
HR4796A is part of a binary star system and has a dwarf companion
about 500 AU away from it. HR4796A was known to have a signicant
infrared excess, and so was a good candidate to have a circumstellar
disc. As in the case of Vega, the temperature inferred from IRAS
observations suggested that the disc did not extend all the way to the
star, but was conned to a ring between about 40 and 200 AU.
The two observing teams reported similar results, both of which
appeared in the same issue of the Astrophysical Journal. The observa-
tions show that HD4796A has an almost edge on disc which extends
out to about 110 AU from the star. The inner edge of the disc cannot be
D
u
s
t

a
n
d

d
i
s
c
s
175
detected on the images directly, but since the dust appears to be at a
temperature of about 110 K, the observers estimated that the inner
boundary was located about 5055 AU from the star. These mid-
infrared observations were soon complemented by observations from
the near-infrared camera NICMOS on the Hubble Space Telescope.
NICMOS had a coronagraphic mode, which made it suitable for
observing faint structures near to bright sources, and its operation at
shorter wavelengths and location above the atmosphere offered much
higher spatial resolution than that available from ground-based mid-
infrared cameras. The NICMOS images showed that the ring was
centred about 70 AU from the star and that it was quite narrow, only
about 14 AU across. The images also conrmed that there was little
material inside about 45 AU.
The outer edge of the dust ring around HR4796A is probably con-
strained by the gravitational effects of the stars companion.
Although small, the companion stars gravity is quite capable of
B
e
y
o
n
d

P
l
u
t
o
176
Figure 10.3 The disc around HR4796A as imaged by the NICMOS camera on the Hubble
Space Telescope. The central region was blocked out by a coronagraph to suppress the
light from the star, revealing the narrow disc. Sophisticated image processing of light
escaping around the edge of the coronagraphic disc was used to reveal structure close to
the centre of the image, resulting in a slightly different appearance of these regions. In
reality, the disc is probably the same all the way around. (Glenn Schnieder.)
stripping away dust which moves too far from HR4796A. However, the
inner edge of the disc cannot be explained in this way. It must be con-
trolled by something interior to the disc. Although the hole might be
cleared by effects such as PoyntingRobertson drag, which would
cause the material to spiral in towards the central star, an obvious
explanation is the presence of one or more planets orbiting inside the
disc. This Planets plus Kuiper Belt model is supported by three other
factors. Firstly, both infrared and sub-millimetre observations suggest
that the mass of dust around HR4796A is somewhere in the region of
one Earth mass, about right for a Kuiper Belt. Secondly, JCMT obser-
vations suggest that there is not much molecular hydrogen gas in the
system, which means that it is no longer possible to form a Jupiter-
sized gas giant planet in the system. If HR4796A is fated to have such
planets, they must already have formed. Thirdly, the MIRLIN observa-
tions point to additional material at temperatures of a few hundred
kelvin, typical of a source within a few astronomical units of the star.
This is just about the location of the bulk of the zodiacal dust in our
own solar system. Taking all these observations together, HR4796A
looks rather like a system mid-way between the massive discs around
young pre-main sequence stars and the more mature systems like
Beta Pic and Epsilon Eridani.
What may well be another extrasolar Kuiper Belt was imaged from
NASAs IRTF telescope by David Trilling and Robert Brown. They
used a cooled coronagraph (CoCo) together with the IRTFs near-
infrared camera to observe 55 Cancri, a Sun-like star believed to be
about 3 billion years old. The existence of a planet a bit larger than
Jupiter orbiting 55 Cancri had already been deduced frommeasure-
ments of tiny wobbles in the stars motion and Trilling and Brown set
out to see what else might be there. By blocking light fromthe star
itself, the coronagraph allowed themto image a region extending as
close as 18 AUto 55 Cancri. Their images revealed an ellipse of scat-
tered light coming froma disc of material extending out at least 40
AUfromthe star. The observations were made through three differ-
ent lters in the 12.5 micron region and the different appearance of
the disc in the different lters gave a clue to the composition of the
material present. In particular, while fairly bright at wavelengths of
1.62 and 2.12 microns the disc was almost invisible at 2.28 microns.
Since methane absorbs light of this latter wavelength, this suggests
quite strongly that the material in the disc has methane ice on its
surface.
D
u
s
t

a
n
d

d
i
s
c
s
177
Trilling and Brown imaged a number of other stars to check that
the extended light around 55 Cancri did not arise from some instru-
mental defect or error in their observing plan. The other stars did not
show any similar features, giving considerable condence that the
excess light around 55 Cancri was indeed due to a Kuiper Belt like disc
of material. Assuming that the disc was roughly circular and at they
were able to deduce, from the shape of the ellipse in their images, that
the disc was tipped about 27 degrees to the plane of the sky. In other
words they were seeing it close to, but not quite, face on. They con-
cluded that 55 Cancri had a mature solar system containing at least
one large planet and a primordial disc.
The SCUBA, MIRLIN, OSCIR and CoCo observations bring us full
circle. They seem to show that dusty and icy discs are common around
main sequence stars and that planet, or at least planetesimal, forma-
tion is a common process throughout the galaxy. The discovery of a
number of other extrasolar planetary systems in recent years has
added to the belief that the solar system is no longer a unique ash in
the pan. It seems that planet formation is a common process which
occurs around many normal stars. It is an interesting conuence of
many apparently unrelated lines of enquiry that has brought together
stellar astronomers and planetary scientists to provide this huge leap
in our understanding of both our own solar system and of planetary
systems around other nearby stars.
B
e
y
o
n
d

P
l
u
t
o
178
Where do we go
from here?
Since it is a subject which is developing so quickly, it is difcult to
predict the future of Kuiper Belt research in detail. None the less
some general directions are clear. Astronomers need better mathe-
matical models of how the Kuiper Belt was formed and evolved, and
they need to understand the chemical composition and physical char-
acteristics of many more individual objects.
To theoretician Martin Duncan the future direction of his work is
clear. He says that physicists have to understand in detail how the
process of planet formation is intertwined with the nal orbits of the
asteroids and comets that we see today. This will require a clearer
picture of how planets accrete from myriads of kilometre-sized plan-
etesimals in the solar nebula. Progress will depend on new and ever
more detailed computer models able to account for, and track, the
mutual interactions of huge numbers of test particles. These new
models will have to include the effects of gas dynamics on the evolu-
tion of the nebula, extra detail which will add greatly to their com-
plexity. Duncan and his ilk will also need increasingly sophisticated
computer algorithms to model the collision and fragmentation of
their test particles. Developing these models will bring together the
very abstract studies of the celestial mechanicians with the practical
work of the snowball smashing experimentalists like Eileen Ryan.
Duncan envisages that ve to ten years from now he and his co-
workers will be using parallel clusters of computers each many times
faster than todays machines. The computers will spend their time
doing fairly detailed simulations of the formation of planets and their
migration across the still-evolving solar system. The results will
reveal how gravitational scattering led to the formation of the Oort
Cloud and will illustrate the gravitational sculpting of the Kuiper
C
H
A
P
T
E
R

1
1
179
Belt. The validity of the models will then be tested by comparison
with the statistics of real comets and Kuiper Belt objects until most,
if not all, of the uncertainties in the models have been removed.
One project which will help to establish some of those observa-
tional constraints is being set up in the mountains of central Taiwan.
The TaiwanAmerican Occultation Survey (TAOS) is an international
collaboration conducting a census of the comet-sized objects in the
Kuiper Belt. Such small objects are too faint to be detected directly, so
TAOS will count them another way. Night after night its telescopes
will monitor a few thousand bright stars and watch for any sudden
dimmings caused as Kuiper Belt objects pass in front of them. Each
occultation will be brief. It will take less than a second for the shadow
of a Kuiper Belt object a few kilometres in diameter to sweep across
one of the TAOS telescope sites. To be sure that the reported dim-
mings are real events and not noise there will be several telescopes set
along a line. Real occultation events will be seen in turn by each of the
telescopes situated along the path of the objects shadow.
Since the detectability of occultations is determined by the bright-
ness of the background star, not by the brightness of the Kuiper Belt
object, the TAOS survey does not require large telescopes. Instead, it
needs to be able to monitor as many stars as possible as frequently as
B
e
y
o
n
d

P
l
u
t
o
180
Figure 11.1 A schematic diagram showing how the TaiwanAmerican Occultation
Survey will make a census of small trans-Neptunian objects. As an object crosses in front
of a distant star it will cast a tiny shadow onto the Earth. If the shadow passes across the
TAOS telescopes, it will be recorded. (Lawrence Livermore National Laboratories.)
possible in order to maximise the chances of seeing the relatively rare
occultation events. The project will start with just three telescopes,
each 0.5 m in diameter, situated in the the Yu-Shan (Jade Mountain)
National Park in Taiwan. Each telescope will be equipped with a 2048
by 2048 pixel CCD camera covering an area of about 1.7 degrees of the
sky. The telescopes will be separated by 510 km along a line that runs
roughly east to west. They must be located sufciently far apart that
chance atmospheric uctuations, clouds and so on, do not affect the
data from more than one telescope, but the spacing cannot be too
large. If the separation is too great then the shadows of the Kuiper
Belt objects, which are unlikely to be moving exactly along the
eastwest direction, will fail to pass over all of the telescopes. If this
happens the events will not be recognised.
The telescope sites are very isolated, so each observatory must
operate automatically. Power will be generated by solar panels during
the day and then stored in batteries until it is needed. Each telescope
will be enclosed by a simple box-like structure topped by a folding
clamshell cover. The cover will fold down out of the way when the tele-
scope is ready to start observing. To protect the telescope against pos-
sible power failures, the cover will be counterweighted and will close
automatically if needed. Once the system is running, the TAOS
project expects to monitor 3000 stars and to produce about 100000
million photometric measurements a year. Out of these anywhere
from a few dozen to a few thousand occultations must be detected,
conrmed by cross checking with the other telescopes in the network
and communicated to the handful of people who are actually running
the experiment. It will be up to them to understand what the results
mean in terms of the population of the trans-Neptunian region.
However, in parallel to modelling how the solar system might be
evolving and studying the statistics of faint Kuiper Belt objects,
astronomers need to be cataloguing more objects to constrain these
very models. Dave Jewitt is very clear about this and speaks elo-
quently for the need to nd many, many more members of the trans-
Neptunian population. One of his favourite themes is to draw an
analogy between the Kuiper Belt and the rather better understood
asteroid belt between Mars and Jupiter. The rst asteroid was found
in 1801 and, within a couple of years, three more were discovered. At
the time these were discovered, almost nothing was known about
their physical properties and no-one really understood what they
were doing there. The rate of asteroid discovery remained slow for
W
h
e
r
e

d
o

w
e

g
o

f
r
o
m

h
e
r
e
?
181
almost a century until new technology, notably the use of photo-
graphic plates, was applied to the search. Once photography was
introduced, many more objects began to be discovered and patterns
in their orbits began to emerge. Gradually, these discoveries led to the
realisation that orbital resonances (mostly connected with nearby
and massive Jupiter) were important in determining the structure of
the present-day asteroid belt. As still more orbits were determined it
was realised that there were clusters of asteroids with very similar
orbits. These clusters are attributed to families of objects formed by
the break-up of larger asteroids, whose fragments are still travelling
around the Sun in much the same orbit as their parent bodies.
Even so, it was not until the 1970s and 1980s that the full complexity
of the main asteroid belt was understood. Only then did a clear
picture of the relationship between asteroids in the main belt and
planet-crossing objects and meteorites emerge. The key to this was
better computers and, most of all, the determination of accurate
orbits for ever more objects. It was these large databases which made
possible detailed statistical studies of the evolution of the asteroid
population as a whole. Together with better ground-based observa-
tions, and recently, images from spacecraft ying past and even orbit-
ing asteroids, the whole complex picture of the history of the asteroid
population was slowly put together.
The observational situation with the Kuiper Belt is very similar to
that of the main belt a hundred or more years ago. Astronomers have
detected only a tiny fraction of the objects that must be out there, and
they have determined reliable orbits for only a few dozen of those.
Just as two centuries ago, new objects are being discovered and then
being lost again due to lack of astrometric follow up. The discoveries
so far suggest that the trans-Neptunian region is far more complex
than anyone ever guessed, and that its structure probably holds
important clues as to how the solar system formed. So, if we are ever
going to understand this region of space, we need reliable orbits for
thousands of objects. This information will make possible detailed
analyses of the populations of the different resonances and of the size
distributions of the objects in each region of the Kuiper Belt. To
Jewitt, the choice is very clear. We can either fool around for another
hundred years or so discovering a few dozen objects a year, or we can
do the job properly, nd and catalogue ten thousand objects in a single
project and solve all of the dynamical problems in one go.
Dave Jewitt believes that what is needed is some kind of dedicated
B
e
y
o
n
d

P
l
u
t
o
182
Kuiper Belt telescope, or at least a sky survey telescope which has
discovery and follow up of trans-Neptunian objects fair and square
amongst its objectives. At a workshop held at the Lowell observatory
in Flagstaff during the autumn of 1998, Jewitt described one such
instrument. He envisaged a 4 m telescope with a eld of view at least
a degree in diameter feeding a mammoth CCD array. Each dark night
the telescope would scan the sky for moving objects, generating
perhaps 100 gigabytes of data each night. Manual searches of this
huge dataset would be impossible, so moving-object software, based on
that already being developed and used by people like Chad Trujillo
and those at Spacewatch, would search the data. The software would
identify likely candidates and pass that information back into the
search programme. This real-time feedback would ensure that subse-
quent observations would be adequate to determine reliable orbits for
each discovery. Nothing would be lost. After a while, the necessary
follow-up observations of hundreds of objects would take a signicant
fraction of the total observing time. Hopefully, within a decade or so,
ten thousand new trans-Neptunian objects would have been found and
tracked in detail. If successful, this effort would reveal the dynamical
structure of the outer solar system once and for all.
It is not clear if such a Kuiper Belt telescope will ever be built, but a
number of studies of large survey telescopes are now underway. One
of these concepts is for a large telescope dedicated to solar system
astronomy to be installed on Mauna Kea. The new telescope would
replace NASAs aging 3.5 m Infrared Telescope Facility. Like the IRTF,
this New Planetary Telescope, or NPT, would be designed to support a
range of NASA projects and would spend most, or all, of its time on
planetary observations. The proposed NPT would be used for studies
of objects being targeted by NASA space missions and to search for
previously unknown asteroids and comets. In particular it would look
for objects which pass close to the Earth and would make a survey of
the trans-Neptunian region. For surveys, the telescope would have to
operate in a wide eld mode, perhaps covering about 4 square degrees
in a single pointing. To detect 10 000 trans-Neptunian objects brighter
than about 24th magnitude would require a search of about 5000
square degrees of sky, something which might be accomplished in a
year or so with the proposed telescope. Astrometric follow-up of these
objects to determine their orbits would require additional time over
the next few years. The time needed for follow-up is why a dedicated
planetary telescope is needed. With so many other interesting
W
h
e
r
e

d
o

w
e

g
o

f
r
o
m

h
e
r
e
?
183
astronomical projects to do, long blocks of time on large telescopes are
generally hard to come by and yet without them much of the effort
spent on searches is wasted.
However, a suitable planetary telescope would not be limited to just
surveying for new trans-Neptunian objects. To accomplish other
aspects of its mission the telescope would have to be designed to have
a very-high-resolution imaging mode, albeit over a much smaller eld
of view. This would mean that it could also be used for studies of indi-
vidual objects. Potential projects might include searches for trans-
Neptunian objects that, like Pluto, are binaries. This would require a
telescope able to go very deep, in order to detect any faint companions,
and to spend considerable amounts of time monitoring the binary
candidates to determine their orbits. However, if binary systems in
the Kuiper Belt can be detected, then the potential benets are enor-
mous. Once the orbital period of the system has been found, then the
masses of the components can be determined quite easily. This in
turn allows the density of the objects to be estimated. The densities
provide important information on the physical make-up of the
objects. For example, it may reveal if the objects are solid or just loose
aggregates of icy boulders. This structural information can then be
related to the outcomes of laboratory impact experiments and mathe-
matical models of the fragmentation process.
At least one other exciting prospect is on the horizon for
astronomers interested in the Kuiper Belt. NASA plans to send a space
mission to Pluto which, if all goes well, will go on to encounter one or
more trans-Neptunian objects. Like most space projects, the mission
has already been through a number of evolutionary changes as
NASAs long-range plans have been modied by budgetary and politi-
cal factors. Although it is still part of the planned NASA programme,
it is by no means impossible that circumstances will force further
changes to the detailed mission prole.
NASA rst considered sending a spacecraft to Pluto as part of a so-
called Grand Tour of the outer solar system. The Grand Tour would
have involved a number of large and complex spacecraft launched
during the 1970s. These ambitious, and very expensive plans were
soon scaled back. They evolved into the highly successful Voyager
project which sent two spacecraft to Jupiter and Saturn. As part of an
extended mission, Voyager 2 went on to explore Uranus and Neptune,
but the ight paths of the two spacecraft were such that neither of
them could be diverted towards Pluto without compromising another
B
e
y
o
n
d

P
l
u
t
o
184
W
h
e
r
e

d
o

w
e

g
o

f
r
o
m

h
e
r
e
?
185
Figure 11.2 A possible conguration for the Pluto Kuiper Express spacecraft. The design
is simple and functional to reduce costs and minimise the likelihood of technical failures.
(Rich Terrile, JPL.)
important scientic objective. Accordingly the outermost planet
remained unvisited. The idea of a mission to distant Pluto then fell
from favour for a while, before resurfacing in the late 1980s.
By 1989 NASA was considering a number of possible new solar
system missions, including ones to explore Neptune and Pluto. These
missions included projects based around the so-called Mariner Mk II
spacecraft, a large multi-purpose craft capable of being used for a
variety of different deep space missions. However, support for space
exploration was dwindling and the NASA budget was shrinking
steadily. The high cost of large planetary missions made the prospects
for a Pluto mission, especially one likely to take twenty years to com-
plete, look grim. So, in 1991, a new concept began to take shape. The
idea originated during a ceremony at NASAs Jet Propulsion
Laboratory to mark the issue of a set of stamps commemorating
Americas planetary exploration projects. All the planets, and the
Earths Moon, were represented with a picture of both the planet and
one of the spacecraft which had visited it. The exception was Pluto,
whose stamp showed just an impression of the planet and the words,
Not yet explored.
The Pluto stamp crystallised the frustration of some of the people
working on proposals for a Mariner Mk II class Pluto mission. Deep
down they knew that a large and complex mission to Pluto would
never be approved and so they began to consider other options. Soon
the idea of a small, low-cost Pluto mission dubbed alternately, Pluto
Very Small or Pluto Fast Flyby started to emerge. The idea was rst
presented to NASA in 1992 and it eventually reached the ofce of
Daniel Goldin, the newly appointed NASA administrator. The pro-
posal was timely, Dan Goldin was trying to move NASA towards a new
fastercheaperbetter philosophy. He believed that the future of
planetary exploration lay in NASA developing numerous small pro-
jects, not in building a few large and expensive spacecraft every
decade. To keep costs down, each of Goldins new generation of
small missions would be aimed at investigating a limited number
of scientic questions on a rapid timescale.
In 1994 NASA decided that even the low-cost Pluto Fast Flyby
was still too expensive and it was decided to try again with an even
cheaper mission. This concept became known as the Pluto Express
and was intended to send two small spacecraft to Pluto. Each space-
craft would have been launched separately and they would begin their
journey by heading towards the giant planet Jupiter. Swinging close
B
e
y
o
n
d

P
l
u
t
o
186
to Jupiter, the spacecraft would use the giant planets gravity to pick
up speed and be ung towards Pluto. They would arrive there, a few
months apart, about 10 years after launch. A scientic denition of
the mission prepared in 1995 described a series of studies to be
carried out at Pluto and mentioned a possible mission extension to y
past a Kuiper Belt object. Given the large number of potential targets
in the trans-Neptunian region, these additional ybys did not present
too many technical difculties. Unless there were severe problems en
route to Pluto, both spacecraft were expected to have sufcient fuel left
to manoeuvre themselves towards suitable Kuiper Belt objects once
their primary mission was complete. Inevitably, the mission contin-
ued to evolve and an early casualty was one of the two spacecraft. A
two-spacecraft mission offered the chance to do additional scientic
observations of the PlutoCharon system and, of course, provided a
safety margin in the event that one of the two craft failed completely.
However, money was tight and NASA could not afford to do everything
it wanted. Faced with a choice between sending two missions to Pluto,
or one mission to each of two different destinations, the decision was
made to restrict the Pluto mission to a single spacecraft and to save
money that could then be used elsewhere.
Although the possibility of an extended mission had been in the
minds of planners for some time, the increasing interest in the Kuiper
Belt meant that eventually the mission was renamed. The mission is
now known as the PlutoKuiper Express, although its primary aim
remains the exploration of the PlutoCharon system. Key scientic
objectives are characterising the surface geology of Pluto and
Charon, mapping the surface composition of both objects and study-
ing Plutos tenuous atmosphere. Although the details of the spacecraft
and its trajectory are likely to change between now and the launch
date, it is already possible to sketch out roughly what the mission
prole will look like.
One scenario developed in the late 1990s envisaged a launch on a
conventional rocket in December 2004, a yby of Jupiter in March
2006 and an encounter with Pluto around Christmas 2012. The space-
craft will be as small and as simple as possible to minimise the risk
of technical problems. The craft will be dominated by a large radio
antenna required to broadcast data back to Earth during and after
the encounter with Pluto. Since the Suns energy is so weak at these
enormous distances, the use of solar panels to generate electricity
is not practical. Instead, power will be supplied by radioisotope
W
h
e
r
e

d
o

w
e

g
o

f
r
o
m

h
e
r
e
?
187
thermoelectric generators similar to these used on other deep space
missions. The PlutoKuiper Express will only carry a few scientic
instruments, probably comprising a visible camera, an infrared
mapping spectrograph and an ultraviolet spectrograph. These should
provide images showing details as small as about 1 km across as the
spacecraft hurtles through the PlutoCharon system. Assuming all
has gone well, the spacecraft may then be directed towards one or
more suitable Kuiper Belt objects. No special equipment will be
carried for this phase of the mission, the scientic instruments
designed for observations of Pluto should be well suited for exploring
any other objects which the spacecraft might encounter.
At present, no specic trans-Neptunian object has been identied
as a possible target. Although Dave Jewitt has conducted a search for
objects close to the expected ightpath of the probe, the region of sky
is close to the Milky Way where the star density makes searching
difcult. He suggests that what is needed is a dedicated search pro-
gramme, perhaps using NASAs proposed New Planetary Telescope,
specically in support of the PlutoKuiper Express mission. In any
event, there is no rush to pick a target as another spanner was soon
thrown into the works. In September 2000 NASA announced that it
planned to expand further its exploration of the planet Mars and soon
afterwards ordered that work on the PlutoKuiper Express be stopped.
Although insisting that the stop-work order was only a postponement,
and not a cancellation, astronomers interested in Pluto fear that the
two things are effectively the same. Even a years delay in the launch
of the Pluto spacecraft will mean that it will no longer be able to use
Jupiters gravity to boost itself towards the outer solar system.
Without such a boost the mission will take longer, or the already small
spacecraft will have to be made smaller. Either way, its arrival at Pluto
will be delayed by years, perhaps to the year 2020. This would be fatal
to many of the missions scientic objectives. Pluto is receding from
the Sun and as it does so it is getting colder and colder. By 2020,
perhaps sooner, the planets tenuous atmosphere will have frozen onto
the surface. Once this happens the atmosphere will remain trapped as
ice for over 200 years. For studies of Plutos atmosphere, time is
rapidly running out.
Inevitably there was an outcry, and attempts to reverse the decision
began. Astronomers argued with NASA while astronomy societies
and individuals lobbied the US congress to have the project re-
instated. However, they may have been too late; NASAs 2001 budget
B
e
y
o
n
d

P
l
u
t
o
188
was approved in November 2000. Attempts to reverse the postpone-
ment will doubtless continue, and hopefully succeed, but only once
the PlutoKuiper express is safely on its way to the launch pad will
searches for its second stop become more important.
W
h
e
r
e

d
o

w
e

g
o

f
r
o
m

h
e
r
e
?
189
Will we ever get
our names
right?
Amongst the many sincere and hotly debated issues concerning the
origin, content and composition of the trans-Neptunian region are a
few questions of little scientic importance. None the less these are
issues of considerable interest. The rst is no less fundamental than
the name of this recently discovered region at the edge of the solar
system; should it really be called the Kuiper Belt, and what should we
call the objects within it?
The termKuiper Belt rst surfaced in the seminal 1988 paper by
Duncan, Quinn and Tremaine in which they reported their conclu-
sions concerning the probable source region of the short-period
comets. They referred to this structure, which at the time was a
purely theoretical concept, as the Kuiper Belt. The name appears to
have been coined by Scott Tremaine, although he says it arose more-
or-less spontaneously since both Kuiper and comet belt appeared
in the opening sentence of Julio Fernandezs 1980 paper on the origin
of the short-period comets. Alan Stern had used the termin 1990
while describing his arguments for an ancient population of ice
dwarfs, but the name attracted little attention until observational
astronomers began to discover real objects in this region. Dave Jewitt
and Jane Luu entitled the paper describing their rst nd as
Discovery of the candidate Kuiper Belt object 1992 QB
1
. In this paper
they spoke quite clearly of the newobject as being the rst known
member of the Kuiper Belt. The name quickly caught on and before
very long all the trans-Neptunian objects were being referred to as
KBOs or Kuiper Belt objects. However, it was not long before some-
body called foul!
Jack Lissauer, from the State University of New York, had men-
tioned Edgeworths 1949 paper when he reviewed the process of planet
C
H
A
P
T
E
R

1
2
191
formation in 1993 and Paul Weissman, a cometary scientist from the
Jet Propulsion Laboratory, had made reference to Edgeworths work
in a review written in 1995. However, the most vocal arguments that
more credit should be given to Kenneth Edgeworth came from
astronomers in Britain. Prominent amongst these were a small
number of people in Northern Ireland who were already familiar with
Edgeworths 1943 and 1949 papers. They, and others, pointed out that
Edgeworth had predicted the existence of a trans-Neptunian disc and
had even remarked that its denizens might enter the inner solar
system to be seen as comets. This, they argued, was as good a predic-
tion as one was likely to get. What was more, Edgeworths papers
came out before Kuiper had published anything on the subject.
Kuipers remarks appeared in 1951, in a chapter which he had con-
tributed to a book edited by J. Allen Hynek.

Surely then, the bodies in

the trans-Neptunian region should be in the Edgeworth Belt, not the
Kuiper Belt? What is more, they noted, Edgeworth spoke of a disc of
material, rather than the ring of material described by Kuiper.
Fuelling the controversy was the fact that Kuipers famous book
chapter made no reference to Edgeworths papers, ignoring the
scientic niceties of citing the work of other workers in the same
eld. Why should this be?
Since both Gerard Kuiper and Kenneth Edgeworth are now dead,
(Edgeworth died in Dublin on 1 October 1972, Kuiper on Christmas
Eve 1973 in Mexico City) we may never know for certain. None the
less, it is interesting to speculate on the circumstances which led to
one of these men being virtually ignored, while the other seems likely
to be immortalised as one of the few people to have not simply a crater,
or a small celestial body, but an entire region of space carrying their
name for the foreseeable future.
Firstly, there is the vexed question of why Kuiper did not refer to
Edgeworths papers in his 1951 review. Could it be that he didnt know
about them? After all, Edgeworths rst paper was published, in
greatly abbreviated form, in the middle of the worst war the world
has ever seen. Not so, says Mark Bailey, director of the Armagh
Observatory in Northern Ireland. Bailey points out that today, with
B
e
y
o
n
d

P
l
u
t
o
192

Hynek was an accomplished astronomer, although is now probably best known for his
categorisation of UFO sightings, one of which was, A Close Encounter of the Third
Kind.
many more scientists active and a plethora of scientic journals and
electronic media in which to publish information, just keeping up
with the ood of scientic literature is a major task in itself. This
was hardly the case fty or sixty years ago. In the 1940s and 1950s
there were far fewer astronomical journals being published and
those which were contained very fewpapers on solar system
research, then considered an astronomical backwater of little inter-
est to astrophysicists as a whole. Kuiper was undoubtedly a world
leader in solar systemastronomy and it would not have been difcult
for himto keep abreast of the small number of papers being pub-
lished in his eld. Even if Edgeworths 1943 note in the Journal of the
British Astronomical Association had somehowslipped past him,
could he have failed to notice the much larger paper in the 1949
Monthly Notices of the Royal Astronomical Society? Dan Green,
writing in the International Comet Quarterly, remarks that in the
1950s the Monthly Notices was in the top three or four of the worlds
most-read astronomical journals and that it would be odd to think
that Kuiper would not be aware of what was being published there.
Dave Jewitt says much the same. Do you think Kuiper wasnt
reading Monthly Notices? he asks rhetorically, He must have been.
So did Kuiper ignore this paper deliberately and somehowsteal
Edgeworths idea? To be fair, Kuiper never claimed the credit for his
prediction. For one thing, he had been dead almost two decades
before the rst trans-Neptunian object was discovered and anyway,
his reputation as a great solar systemastronomer was already
secure.
Gerrit Pieter Kuiper was born in the Netherlands on the 7th of
December 1905. As a young man, he was an outstanding student whose
astronomical interests were encouraged by his father and grand-
father. It was they who gave him a small telescope with which to
pursue his hobby. Kuiper was awarded a degree from the University of
Leiden in 1927 and entered postgraduate studies immediately after-
wards. His professors at Leiden included such famous astronomers as
Ejnar Hertzsprung, Willem de Sitter, Jan Woltjer and Jan Oort.
Kuipers PhD thesis concerned binary stars and in 1933, his thesis
complete, he moved to the Lick Observatory in California. After a
brief stay at Lick he moved by way of Harvard University to the
Yerkes Observatory, part of the University of Chicago. He continued
to work in the area of stellar astronomy for many years.
W
i
l
l

w
e

e
v
e
r

g
e
t

o
u
r

n
a
m
e
s

r
i
g
h
t
?
193
In the winter of 19431944 Kuiper turned his attention briey onto
the planets. He found the rst evidence that Saturns satellite Titan
has an atmosphere containing methane gas and this discovery
diverted him to solar system research. Using new infrared detector
technology developed during the Second World War he began a pro-
gramme of infrared spectroscopy of the giant planets and their satel-
lites. In 1948 he discovered Miranda, the fth known satellite of
Uranus and in 1949 he found Nereid, the second moon of Neptune. He
left the Yerkes Observatory in 1960 and founded the Lunar and
Planetary Laboratory in Tucson, Arizona, still a major research insti-
tute today. In the 1960s Kuiper worked on various NASA space pro-
jects, including the robotic Ranger and Surveyor lunar missions. He
had few research students, claiming he was too busy to supervise
them, but those who did study with him included some very accom-
plished planetary scientists of the next generation, including Tom
Gehrels, William Hartmann, Carl Sagan, Dale Cruikshank and Toby
Owen.
In parallel with his research projects, Kuiper was always inter-
ested in nding better sites for ground-based observations and was
instrumental in establishing major observatories in both Chile and
Hawaii. He was also in the forefront of developing infrared telescopes
carried in high-ying aircraft. As a tribute to his work in so many
areas of solar systemastronomy, Kuipers name is carried on one of
the brightest craters on the planet Mercury, a large crater on Mars, on
Minor Planet 1776 Kuiper and by NASAs Kuiper Airborne
Observatory.
Although some people believe Kuiper had a bad habit of minimis-
ing the work of other people by making very selective use of refer-
ences, scientists who knew him, including ex-students of his still
active in planetary astronomy, offer other possible explanations for
his failure to refer to Edgeworths work. Firstly, they point out that the
social structure of astronomy has changed in the last fty years.
Today, scientists often work in large teams, but then scientic papers
with several authors were quite rare. Similarly, today most papers
have comprehensive lists of references pointing to work done by other
W
i
l
l

w
e

e
v
e
r

g
e
t

o
u
r

n
a
m
e
s

r
i
g
h
t
?
195
Figure 12.1 Gerard Kuiper (with his belt) stands outside the Yerkes Observatory in the
1950s. (Yerkes Observatory.)
people in the eld,

but this was not always the case. Secondly,

Kuipers interests spread across the gamut of problems in solar
system astronomy. While he may have seen Edgeworths papers, he
may have already come to much the same conclusions himself, or have
subliminally absorbed them and then later thought of them as mostly
his own. According to Dale Cruikshank,

Kuiper was a strong person-

ality whose native brilliance was accompanied by an uncompromis-
ing assurance of the importance of his own work. Cruikshank thinks
it highly likely that Kuiper knew about Edgeworths papers, but that
he just ignored them and developed the same ideas himself. This is an
opinion supported by astronomer Willem Luyten, who knew Kuiper at
the Lick Observatory. Luyten, writing in 1979, said that he had no
doubts about Kuipers intelligence, but felt that Kuiper was sometimes
able to put completely out of his mind that somebody else had had an
idea rst and would simply adopt it as his own. Edgeworths case may
not be the only example. According to Luyten, a Dutch astronomer
named H. P. Berlage, whom he had known years before as a fellow
student, came to him in 1946 with a voluminous paper describing a
model for the formation of the solar system from a turbulent solar
nebula. Although Luyten warned his old acquaintance that the paper
was too long, Berlage had no time to modify his manuscript. It was
submitted unchanged to the Astrophysical Journal, whose editorial
ofces were at the Yerkes Observatory where Kuiper was the director.
As Luyten had feared, the paper was not accepted and was eventually
returned unpublished. However, Kuipers chapter in Hyneks 1951
book described a theory for the origin of the solar system which
closely paralleled Berlages ideas. It is by no means impossible that
Kuiper had seen Berlages manuscript while it was at Yerkes.
Kuiper also had a reputation of never quoting the work of another
scientist unless he knew them personally and had a good grasp of
their abilities. Edgeworth was an unknown quantity, an amateur theo-
retical astronomer working in a foreign country. Perhaps Kuiper
somehow didnt feel it was appropriate, or relevant, to make reference
to his work. There is also a question of timing. Edgeworths paper was
received by the Royal Astronomical Society in June 1949 and appeared
B
e
y
o
n
d

P
l
u
t
o
196

Whether this is to give due credit to other people, or merely to minimise the risk of
offending potential referees who must approve the paper for publication is not always
clear.

To whom I am indebted for a copy of his excellent biographical sketch of Kuiper origi-
nally published in the Biographical Memoirs of the National Academy of Science.
in one of the last issues of that years journal. However, the issue was
not actually printed until March of 1950. So it might not have reached
the Yerkes Observatory until the late spring or early summer of 1950.
Kuiper later claimed that the bulk of his chapter in Hyneks 1951 book
was done as early as 1949, which would mean that it was written about
the same time as Edgeworths paper. However, Kuipers chapter does
include references to, and discussion of, the 1950 papers of Jan Oort
and Fred Whipple. There is also a citation of one of Kuipers own 1951
papers so it seems that at least some late updating had been done
before the book went to press.
Kuipers review, being in many other ways so comprehensive, may
have distracted other workers away from the earlier literature. Julio
Fernandez, who turned his attention to the problem of the short-
period comets in the late 1970s, admits he never saw any of
Edgeworths papers while he was doing his research. Of course, by
then Edgeworths work had been forgotten for almost twenty years
and Fernandez can perhaps be forgiven for not searching so far back
into the literature. However, Kuiper and Fernandez were not alone in
passing over Edgeworths writings. By 1961, Edgeworth had written
an entire book, The Earth, The Planets and the Stars: Their Birth and
Evolution, which was published by the MacMillian Company in New
York, but the very next year Alastair Cameron wrote a major review
of theories of the origin of the solar system without mentioning
Edgeworth. Cameron did however give at least passing mention to
Kuipers remarks about a trans-Neptunian comet population.
Cameron was a nuclear physicist who had been working on problems
associated with how elements are built up inside stars and his main
interest in Kuipers work was to do with Kuipers treatment of gas
dynamics in the outer solar system. He was not particularly inter-
ested in the orbital distribution of comets and at the time had never
heard of Kenneth Edgeworth. Looking back across a gulf of forty
years, Cameron thinks that even if he had known about Edgeworths
book, its title would have suggested to him that it was a popular work
and he would probably not have pursued it.
What about the British astronomical establishment? Surely they
could hardly have failed to notice a paper in their most prestigious
astronomical journal? Yet here too, Edgeworths work sank into
obscurity. Again we can only speculate, but perhaps the professional
astronomers of the time did not like the idea of someone muscling in
on their territory. This might have been especially true of someone
W
i
l
l

w
e

e
v
e
r

g
e
t

o
u
r

n
a
m
e
s

r
i
g
h
t
?
197
who lacked a university education or a position in a recognised astro-
nomical research group. The strong bonds that today link amateur
astronomers, especially those working in cometary observing, with
their professional counterparts were almost non-existent sixty years
ago. Perhaps Edgeworths writings were seen by the academic world
as little more than irrelevant scribblings produced by someone on the
fringes of real astronomy.
Another possibility is that for Edgeworth, the timing of his papers
was just a bit too early. Edgeworth had described the objects in his
putative disc as clusters, echoing then prevalent theories that comets
were loose aggregations of particles, ying gravel banks so to speak.
His paper appeared just before the now popular icy conglomerate
model gained acceptance. So, in 1949, there was no reason to believe in
solid icy objects in the outer solar system. However, Kuiper, when he
wrote his review a little later, had the advantage of knowing about the
work of Fred Whipple and Jan Oort, which was revolutionising the
understanding of comets. Even so, according to Dan Green, it seems
that Kuiper thought of the trans-Neptunian region mostly as a place
where comets had been formed early on before being ejected by Plutos
gravity into the Oort Cloud. Kuiper does not appear to have consid-
ered that a trans-Neptunian reservoir of comets might exist today.
Edgeworth certainly seems to have lost interest in the astronomical
establishment at some point. Although he was a visitor to various
observatories in Ireland, and lectured to the Irish Astronomical
Society about the origin of the solar system on 9th April 1951, he did
not publish much further work after 1949. His later works were
limited to his 1961 book and a couple of letters in The Observatory
magazine. Despite his obvious interest in the subject, Edgeworth
made little reference to astronomy in his autobiography, Jack of all
Trades, the Story of my Life, which was published privately in 1965. In
just two paragraphs, he merely noted his long-standing interest in the
subject and that he had published a few papers and a book concerning
Certain problems in astronomy. That these papers deserved at least
some consideration was eventually noted by the naming of Minor
Planet 3487 Edgeworth in May 1999.
Whatever the merits of a case for posthumous recognition of
Edgeworths work, it may already be a losing battle in which a few
people are simply ghting a desperate retreat. Astronomy is full of
examples of things that got named after the wrong person and
Edgeworth was not the rst, and is unlikely to be the last, to lose out in
B
e
y
o
n
d

P
l
u
t
o
198
this way. A classical example of this mis-crediting is probably the so-
called Bodes Law. This set out a numerical sequence which, for a time
at least, seemed to explain the relative distances of the various planets
from the Sun. The relationship was actually discovered by Johann
Titius von Wittenberg; it was merely publicised by the far better
known, and so more inuential, Johann Bode. In this regard, one
could also point to the naming of the Oort Cloud. Although Oorts
name is correctly associated with the idea of a distant comet cloud,
many of his conclusions were presaged by Ernst Opik in 1932 and
fellow Dutchman Van Woerkom in 1948. Dave Jewitt admits he has
now become more aware of the work of Edgeworth and thinks that he,
not Kuiper, deserves the credit for the prediction. Martin Duncan con-
cedes that he doesnt know what his group would have done about a
name for the comet belt had they known about Edgeworths papers in
1988. None the less, the term Kuiper Belt has already entered common
usage. Experience suggests that it is unlikely to be formally revoked
or fall rapidly from grace.
In an attempt to rectify the perceived injustice of this situation,
astronomers have tried to nd a solution that respects both parties,
and some refer to the region as the EdgeworthKuiper Belt. Others
say, in effect, that since like it or not we have a Kuiper Belt, why not ll
it with EdgeworthKuiper Objects, or EKOs? Some will not even go
this far. Brian Marsden, for one, counters that neither Kuiper nor
Edgeworth really got it right. It was, he says, Fred Whipple who rst
wrote realistically and quantitatively about a population of comet-
like objects in the trans-Neptunian region. According to Marsden,
Whipple sketched a comet belt extending from about 3550 AU which
stopped rather abruptly. Whipple said These things are about a
100 km in diameter and have V magnitudes of about 22. We have no
hope of detecting them at present. Marsden argues that this is closer
to what we are actually talking about today and says, Neither
Edgeworth or Kuiper wrote about anything remotely approximating
to what we are seeing, but Fred Whipple did. Julio Fernandez also
credits Whipple with helping to disseminate ideas about a comet belt.
Fernandez describes a paper written by Whipple in 1972 as being very
inuential in his own thinking about the trans-Neptunian region.
To further confuse the issue, Dutch journalist George Beekman
added a new angle to the debate in 1999. Writing in the October issue
of the Dutch magazine Zenit, he reported that Armin Otto Leuschner
was quoted in the 14th April 1930 issue of the New York Times as
W
i
l
l

w
e

e
v
e
r

g
e
t

o
u
r

n
a
m
e
s

r
i
g
h
t
?
199
speaking of Pluto as being possibly a bright cometary object.
Professor Leuschner was an astronomer at the University of
California and he numbered amongst his students Fred Whipple and
E. C. Bower. Based on the scanty information being released from the
Lowell Observatory, Leuschner and his students had computed no less
than six possible orbits for Pluto. In his remarks to the press,
Leuschner seemed to have hedged his bets, remarking that Pluto
might be an escaped main belt asteroid ung into the outer solar
system by Jupiter, a long-period planetary object or a giant comet.
Leuschner does not seem to have pursued the idea of a trans-
Neptunian belt and, according to Marsden, Fred Whipple has no recol-
lection of any discussion with Leuschner on the subject of a
multiplicity of objects beyond Neptune. Marsden has however drawn
attention to remarks by F.C. Leonard in an August 1930 leaet pub-
lished by the Astronomical Society of the Pacic. Leonard wrote,
Now that a body of the evident dimensions and mass of Pluto has
been revealed, is there any reason to suppose that there are not other,
B
e
y
o
n
d

P
l
u
t
o
200
Figure 12.2 A sketch of a hypothetical comet belt published by Fred Whipple in 1964.
Extending from 3550 AU, Whipples prediction was very close to what is now known as the
Kuiper Belt. (F. Whipple.)
probably similarly constituted, members revolving around the Sun
outside the orbit of Neptune?. Leonard continued, Is it not likely that
in Pluto there has come to light the rst of a series of ultra-Neptunian
bodies the remaining members of which still await discovery, but
which are destined eventually to be detected?. If this was a guess,
it was a spectacularly good one.
Mark Bailey, one of Edgeworths champions, raises another
warning. He notes that since 1993 the expression Kuiper Belt has
come to cover a wide range of things including objects in resonances,
those in the classical belt and those in the scattered disc. He cautions
that, We should be careful about blurring important distinctions and
try to use words with precision.
Before moving on to more emotive, and even less scientically pro-
ductive issues, let us address one last question on the subject. Is it a
belt, or a disc? Alan Stern much prefers to call it a disc. After all, he
argues, the region is extended both radially outwards from the Sun
and above and below the ecliptic plane. That is a disc, not a belt.
Although Sterns argument is sound, it falls victim to the same
problem of usage as before. The early papers of Duncan, Jewitt and
others refer to the structure as a belt, and a belt it looks like staying.
According to Hal Levison, Stern bowed to the inevitable when it was
pointed out to him that if he continued to put Kuiper Disc, rather than
Kuiper Belt, in the titles of his papers then computerised search
engines were in danger of failing to nd his work. If this were to
happen, the citation rates of his papers would be reduced as no-one
would ever read them. So, like it or not, it looks as if astronomy is
stuck with a Kuiper Belt, albeit a thick one containing EKOs.
And what of these EKOs, what shall we call them individually?
Traditionally, minor planet names have been suggested by the objects
discoverer after observations spanning two or three oppositions have
been made. By then the orbit is usually well-enough dened that the
object is not likely to be lost again. The name is then approved by the
Small Bodies Names Committee of the IAU. The Minor Planet Center
uses the probable error in an objects orbit to determine when it might
become ready to be named, but this scheme becomes more difcult
when dealing with objects at great distances from the Sun. The slow
motions and small observed arcs of these very-distant objects mean
that it will be some time before they are going mathematically to
reach the status of numberable minor planets (although the orbits
for many of them are now quite secure) and when they do, it is quite
W
i
l
l

w
e

e
v
e
r

g
e
t

o
u
r

n
a
m
e
s

r
i
g
h
t
?
201
possible that at least some of the original discovers might not even be
alive to name them. There is also the question of their minor planet
numbers. Should they be simply slotted into the regular list of minor
planets? If this is done then there will be no way to distinguish a faint
main belt asteroid from a faint EdgeworthKuiper object. A new
system, using a new series of numbers prexed with K was suggested,
but this idea did not prove popular, and as of spring 2000, no Kuiper
Belt object had been numbered, let alone named ofcially.
However, a few objects have been given unofcial names. None of
these has yet come into general use, which in some cases is perhaps
just as well. For a time, the Spacewatch team called the Centaur they
discovered in 1992 Big Red. It is not a bad name, the object is quite
big, and it is certainly red, but such a sobriquet was never going to
catch on ofcially. The object was subsequently named Pholus.
Centaur 2060 Chiron has had its ups and downs as well. Although
a few people had suggested quite early on that Chiron was an inert
comet, the absence of any actual cometary activity around the time of
its discovery meant that it was originally classied as an asteroid. The
onset of cometary activity a few years later presented something of a
problem. If Chiron was a comet, then perhaps it should have a comets
name and be known as Comet P/Kowal. This dilemma was solved by
Mike AHearn, a respected cometary astronomer from the University
of Maryland, who suggested that the object have some kind of dual
status. As a result, 2060 Chiron is also designated 95P/Chiron, the 95P
denoting that it is also the 95th comet proven to be periodic. How
astronomers actually refer to the object often depends on what aspect
of its behaviour they are investigating. Sometimes it is comet
95P/Chiron and other times just plain old Chiron.
The saga of unofcial names also extends to the rst few trans-
Neptunian objects. Soon after its discovery, news circulated that 1992
QB
1
was being called Smiley. There was a grain of truth in this.
Smiley was indeed the name Jewitt and Luu had given to the object
when they found it. Luu was reading one of the spy novels of John Le
Carre and, needing to call their discovery something for their own
book-keeping purposes, they picked Smiley, Le Carres spymaster. It
seemed a reasonable name for something that had remained hidden
for a long time and was only tracked down after much effort.
Following the logic of this, they called the second object (1993 FW)
Karla, codename of Smileys main opponent. Although Luu says that
the names were only intended for their own use as they tried to keep
B
e
y
o
n
d

P
l
u
t
o
202
track of their various candidate objects, the more conservative ele-
ments in the astronomical community were not amused. It was very
quickly pointed out that there was already an asteroid called Smiley.
It is Minor Planet 1613, an otherwise obscure main belt object about
22 km in diameter which was discovered in 1950. It was named after
Charles Hugh Smiley, a celestial mechanician who died in 1969. There
is also a Minor Planet 1470 Carla and this similarity would more-or-
less rule out calling something Karla. However, Smiley and Karla were
never formally proposed to the IAU and when asked about it some
time ago Dave Jewitt said, Lets not talk about names. Jane Luu
conrmed that in her opinion, There are better things to worry about
than the names.
However, Smiley and Karla would probably have been quite accept-
able to the IAU compared with the informal names being used by
Williams, Fitzsimmons and company when they discovered 1993 SB
and 1993 SC. Talking around the subject at the observatory cafeteria,
over what was probably one glass of wine too many, they reasoned
that since Pluto was the name of a cartoon dog, the objects past Pluto
might logically be named after cartoon cats. They soon abandoned the
idea, but for a while the outer solar system was populated, unofcially
at least, not just by Smiley and Karla, but by Felix and Gareld as well.
While the naming of individual objects is a matter for the discover-
ers and the appropriate bodies of the IAU, and the choice between
Kuiper Belt, or Edgeworth Sheet, or even Duncans Donut, will most
probably be made by the custom and practice of astronomers active in
the eld, there is one naming issue which certainly caught the atten-
tion of the press and public. It concerned the planetary status of no
less an object than Pluto.
The issue arose when it became clear that Pluto was just one of
a number of objects in the 2:3 Neptune mean motion resonance. In
many respects, Plutos orbit is quite indistinguishable from those of
other Plutinos and some astronomers soon began to refer to Pluto as
simply the largest of the known objects in the Kuiper Belt. Brian
Marsden had remarked on this as early as 1992 when discussing the
description of 1992 QB
1
as the rst Kuiper Belt object. He told the
Boston Globe that, It was probably unfortunate that Pluto has been
considered a planet, and asked, Is 1992 QB
1
the rst Kuiper Belt
object, or is Pluto the rst?. Things began to heat up when the
increasing rate of discoveries from automated search telescopes
meant that the number of numbered minor planets was rapidly
W
i
l
l

w
e

e
v
e
r

g
e
t

o
u
r

n
a
m
e
s

r
i
g
h
t
?
203
heading towards 10 000. Since there was a tradition of naming aster-
oids with round numbers after someone or something special, such
as 1000 Piazzi, 2000 Herschel and so on,

Brian Marsden suggested that

Pluto might be numbered as minor planet 10 000. This would recognize
Plutos status as part of the Kuiper Belt. Marsden discussed the issue
with members of the appropriate naming committee who agreed that
this idea was viable. Mike AHearn, who had conceived the dual status
idea for Chiron a few years earlier, was in favour of Pluto having dual
status too. The idea of numbering Pluto soon began to circulate
within the wider solar system community.
It was not a new issue. The question of Plutos status as a real
planet had come up from time to time before, especially when it was
realised that the planet was much smaller than rst thought. However,
the issue had never really come to a head, perhaps because until the
minor planet 10 000 idea surfaced, no-one had any better ideas.
Marsden insists he never suggested demoting Pluto from the list of
planets, suggesting merely that giving it a minor planet number
would make things tidier when the other trans-Neptunians began to
be numbered and catalogued. The dual status idea was seen by him as
a compromise between the physical reality of Pluto as a large trans-
Neptunian object and its traditional identity as a planet. Despite his
good intentions, the proposal ignited a restorm of criticism which,
like most such debates, generated more heat then light.
On the one hand stood the group in favour of numbering Pluto.
Some of their arguments ran as follows. Pluto is orbitally indistin-
guishable from the other objects in the 2:3 resonance. Indeed, if you
plot the objects in orbital element space, only an expert could pick
Pluto out from the other objects in resonance with Neptune. Orbitally
speaking, there is just nothing special about it. Although he would not
go as far as saying Pluto should be demoted, as a dynamicist, Martin
Duncan comes down rmly on the side of Pluto as a large
EdgeworthKuiper object. Hal Levison agrees, saying, I rmly
believe that if Pluto were discovered today we wouldnt be calling it a
planet. This sentiment is echoed by Brian Marsden, who rules that if
discovered today Pluto would have got a temporary designation, then
a minor planet number, just like any other similar object.
B
e
y
o
n
d

P
l
u
t
o
204

The others are 3000 Leonardo, 4000 Hipparcus, 5000 IAU, 6000 United Nations, 7000 Curie
and 8000 Isaac Newton. Minor Planet 9000 broke with tradition a bit, it was called Hal
after the talking computer in the lm 2001: A Space Odessy.
Next comes the issue of Plutos size. When it was rst discovered
Pluto was thought to be much bigger than we now know it to be.
Indeed, the estimated size of Pluto has shrunk steadily for many
decades and a light-hearted paper published in the 1980s showed that
the decrease in the estimated size of Pluto could be tted by a curve
that would predict the planets complete disappearance quite soon. On
the question of size, Marsden says, If you want to consider a planet as
something that is spherical and has collapsed under gravity, then you
can have many such objects, fteen, twenty, maybe more. There are
plenty of main belt asteroids which would meet these standards.. He
describes the original naming of Pluto as a planet as irrational and
feels that it only came about because the Lowell observatory did a very
good public relations job of saying that they had found the planet pre-
dicted by Percival Lowell. Of course, we now know that Pluto was not
Lowells Planet X. Pluto was found by chance, or rather by reason of
the very careful search that Tombaugh had made.
On the other side of the divide stood the traditionalists who
argued, often passionately, that Plutos status as a planet should not be
imperilled. It is clear that the Lowell observatory remains very proud
of its planet and Robert Millis, director of the observatory, says that
numbering Pluto makes, No sense, remarking that Pluto deserves to
be considered as more than a minor planet. Alan Stern, a staunch
defender of Plutos planetary status, agrees. He says that, Its not as if
the Lowell staff knew Pluto was only 1000 km or so in size when they
found it and they were pushing it on people. They really thought they
had discovered a new planet. Whatever we know now, the planetary
lobby say, Pluto has been classied as a planet for over 60 years and to
reclassify it now would be foolish and a break with tradition. Jim
Scotti of Spacewatch agrees, People categorise things to bring some
kind of order to things and Pluto was classied as a planet so lets
keep it that way. Marsden counters that this argument is not strictly
correct. What we now call Minor Planet 1 Ceres was once happily des-
ignated the eighth planet (Neptune had not yet been discovered) and
the next few asteroids were regarded as planets too. It was only with
the discovery of more and more asteroids that Ceres, Vesta, Pallas,
Juno were reclassied as a minor planets. Indeed, astronomy text-
books published as late as 1847 referred to eleven primary planets.
Continuing the defence of Plutos planetary status are those who
point out that Pluto is different from the rest of the objects in the 2:3
resonance in several important ways. Firstly, it is much bigger; with
W
i
l
l

w
e

e
v
e
r

g
e
t

o
u
r

n
a
m
e
s

r
i
g
h
t
?
205
a now well-established diameter of 2300 km and a mass 0.00237 times
that of the Earth. Pluto is small in planetary terms, but still several
times bigger than its nearest rival amongst the Plutinos. The com-
pleteness of the searches by Tombaugh and Kowal makes it very
unlikely that a brighter Plutino will ever be found. Secondly, Pluto has
a satellite, and no other trans-Neptunian object has one. This is no
longer a clinching argument as some main belt asteroids such as 243
Ida are now known to have tiny satellites, but Plutos satellite Charon
is large. In fact, Charon is so large compared with Pluto that many
astronomers refer to the system as the PlutoCharon binary regard-
ing it as a sort of double planet rather than an ordinary planet plus a
moon. Finally, Pluto has an atmosphere, albeit a thin one that will
freeze out on the surface sometime in the early decades of the twenty-
rst century as Pluto recedes from the Sun. So there you have it say
the traditionalists, Pluto is spherical, bigger than the other things
around it, has a moon and an atmosphere. What more do you want to
make it a planet? Alan Fitzsimmons is pro-planet, although he admits
that Pluto is a bit of both. Schizophrenic is the word he used.
The issue rumbled on for a while with Brian Marsden urging that
Pluto get the coveted minor planet number 10 000, but overall senti-
ment was against him. As the debate spread, more and more
astronomers pitched in. David Hughes from England said that, It was
just astronomers admitting to what they have known for a long time,
while others claimed demoting Pluto was stupid. Soon the debate
started to spread outside the tight-knit community of solar system
astronomers via the press and internet, becoming simultaneously
more heated and less informed as it did so. Eventually, the General
Secretary of the IAU, Johannes Anderson, felt obliged to make a
formal statement of the IAUs position. Although thought to be per-
sonally in favour of numbering Pluto, on 3rd February 1999 Anderson
stated that, No proposal to change the status of Pluto as the ninth
B
e
y
o
n
d

P
l
u
t
o
206
Figure 12.3 Pluto and Charon. Kuiper Belt objects or double planet? (Gemini
Observatory.)
planet in the solar system has been made by any Division,
Commission or Working Group of the IAU responsible for solar
system science.
Despite this announcement, various attempts were made to mount
a vote on the issue. Many people felt this was not a sensible way
forward. Alan Stern remarked, If we were to take a vote and rename
Brian [Marsden] an amateur astronomer it wouldnt change the
things he has done in his career, but it would be pejorative. It would
taint things. Marsden says he never meant to be pejorative. He thinks
that Pluto is a most interesting object, but that its just not a planet.
None the less, votes were taken. One conducted by the Minor Planet
Center came out strongly in favour of numbering Pluto, although an
informal poll taken at a meeting of asteroid astronomers in Germany
during 1998 came out with the opposite result.

According to Mike
AHearn the debates have been remarkably emotional and the most
interesting conclusion of the whole affair is the demonstration that
astronomers are less rational than he thought they were.
Part, indeed most, of the problem is that there is no formal
denition of a planet. Furthermore, it is very difcult to invent one
which would allow the solar system to contain nine planets. Alan
Stern feels that moving away from the rhetoric and actually dening
what makes something a planet would help to crystallise our think-
ing. He suggests that for an object to be classied as a planet requires
it to have three characteristics. It must be in orbit around a star (thus
removing the larger satellites from contention), it must be too small to
generate heat by nuclear fusion (so brown dwarf stars are excluded)
and it must be large enough to have collapsed to a more-or-less spheri-
cal shape (which excludes comets, and most of the asteroids). These
criteria would admit a few of the larger asteroids and probably some
of the Kuiper Belt objects as well, but adding a requirement for a
planet to have a minimum diameter of 1000 km would remove the
larger asteroids from contention while retaining Pluto. However,
setting a diameter of 1000 km is very arbitrary (why not use 1000
miles?) and it has no physical meaning in terms of how the objects
formed or evolved. After all, if Plutos companion Charon was just a
bit larger, would it be called a satellite, or fully conrmed as half of a
binary planet?
W
i
l
l

w
e

e
v
e
r

g
e
t

o
u
r

n
a
m
e
s

r
i
g
h
t
?
207

The result from the German poll was 20 to 14 with a lot of abstentions.
Astronomers catalogue things to try and rationalise the Universe
and to understand how it works. Objects which fall on some border-
line, like the comet-cum-asteroid Chiron, can be studied by people
with different perspectives and often provide crucial tests for cher-
ished theories. From the point of view of the evolution of the solar
system, it is sensible to consider Pluto along with the other trans-
Neptunians, but from the point of view of how large solid bodies
behave internally, Pluto is best considered a small planet. What the
whole debate may be telling us is that there are at least three types of
planets; rocky terrestrial planets like the Earth and Mars, giant
planets like Jupiter and Neptune and a recently recognised class of
ice dwarfs which encompasses Pluto, Charon, some of the large icy
satellites and the large trans-Neptunian objects.
The arguments eventually subsided due to a mixture of scientic
realism, tradition and perhaps an eye towards astronomys public
relations. From a scientic point of view it doesnt matter what we
call Pluto. Reclassifying it wont actually help us understand its com-
position, origin or future. Jane Luu says, Sure, we can call it a planet,
its no skin off my back, as if to dismiss the issue as trivial compared
with actually studying the object itself. Jewitt is rmly in the camp of
those who regard Pluto as a large EKO, but says he is happy if people
want to call it a planet. People are confused about lots of things, he
says. Julio Fernandez is also in favour of the status quo. Astronomy is
full of things with names that later proved to be incorrect or just plain
silly, he points out. The lunar Maria, or seas, were named when
astronomers thought they were just that, large bodies of open water.
Three hundred years later, when Neil Armstrong made his small step
onto the moon he was wearing boots, not ippers, but no-one has sug-
gested renaming the Maria. Probably no-one ever will. One other
remark comes from Eileen Ryan who once shared an ofce with Clyde
Tombaugh and who favours the status quo. Every time the subject
came up Clyde was just crushed she said.
Finally, there was the issue of the public perception of astronomy.
Science progresses by continually investigating and challenging
existing beliefs and sometimes nding out that earlier ideas were not
quite as good as once thought. The idea that the Earth was the centre
of the Universe held sway for centuries until a better idea came along.
Even though it was correct to do so, demoting the Earth from its privi-
leged position took some doing. Even in todays more rational times,
for some people giving up an idea held dear for many years is difcult.
B
e
y
o
n
d

P
l
u
t
o
208
It seemed to some that there was a real risk that a media interested in
scandal and controversy would distort the issue and imply that the
scientists had somehow, Got it wrong again. Although Mark Bailey
described the issue as a great opportunity to educate people about
science, and Dan Green said it was time to stop teaching our children
an outdated 1940s picture of the solar system, at least a few
astronomers feared demands for the withdrawal of textbooks and pre-
sumably the ring of the incompetents who could not tell what a
planet looked like.
The witching hour came and went when minor planet number
10000 was nally reached. It was assigned without fanfare to a small
main belt asteroid which was named Myriostos, which means 10 000th
in Greek. Of course astronomers may live to regret this decision if an
object bigger than Pluto is ever found amongst the classical Kuiper
Belt or, more likely, in the scattered disc or the Oort Cloud. Certainly
there is no reason why a few large Kuiper Belt objects might not exist
in the dark outer reaches of the solar system. The gravitational forces
which eject objects into the disc do not care how big they are, as long
as they are small compared with one of the giant planets. So a Pluto-
sized, or even-Earth sized object may well be lurking out in deep space
waiting to be discovered. If one ever is, then Brian Marsden may have
the last laugh. Perhaps he even has a bet on it with someone.
W
i
l
l

w
e

e
v
e
r

g
e
t

o
u
r

n
a
m
e
s

r
i
g
h
t
?
209
Dramatis
personae
Manyof the astronomers mentionedinthis storyhave beenhonoured
byhavingaminor planet namedafter them. The citations for those
minor planets are givenhere. Theywere true at the time of the naming,
but inat least some cases theyare nolonger uptodate. Topreserve the
avour of the original citations, I have not attemptedtoupdate them. It
is important torecognise that minor planet namingis not asystematic
process andsothis list is bynomeans comprehensive. There are cer-
tainlyastronomers whohave contributedmightlytothis eldandnot
yet beenassuredof aplace inthe heavens. Here are afewwhohave.
(3192) AHearn
1982 BY
1
. Discovered 1982 January 30 by E. Bowell at Anderson Mesa
Named in honor of Michael F. AHearn, professor of astronomy at the
University of Maryland. A prominent student of cometary physics,
AHearn has pursued coordinated spectroscopic and spectrophotomet-
ric observations of comets spanning the spectral interval from the
vacuum ultraviolet to the radio region. He participated in the 1983 dis-
covery with the IUE spacecraft of diatomic sulphur in the spectrum of
Comet IRASArakiAlcock (1983d) and has made many other impor-
tant contributions to our current understanding of comets.
(4050) Mebailey
1976 SF. Discovered 1976 September 20 by C.-I. Lagerkvist and H.
Rickman at Kvistaberg
Named in honor of Mark E. Bailey, a British astronomer at the
University of Manchester well known for his work on the origin of
A
P
P
E
N
D
I
X

1
211
comets, the dynamics of the Oort Cloud and the capture of comets
into short-period orbits.
(3485) Barucci
1983 NU. Discovered 1983 July 11 by E. Bowell at Anderson Mesa
Named in honor of M. Antonietta Barucci, planetary scientist at the
Istituto di Astrosica Spaziale in Rome. A prolic contributor to the
study of the physical properties of minor planets, Barucci has carried
out both photometric and astrometric observations at the telescope
and has studied minor planet body shapes and surface light-scattering
properties in the laboratory.
(7553) Buie
1981 FG. Discovered 1981 March 30 by E. Bowell at Anderson Mesa
Named in honor of Marc W. Buie (1958 ), an astronomer at Lowell
observatory, who has made many important contributions to plane-
tary astronomy. These include the identication of water ice on
Plutos satellite Charon, constraints on the albedo and frost distribu-
tion on Pluto and Charon and improvements in our knowledge of
Charons orbit. Buie is also a codiscoverer of several trans-Neptunian
objects and has developed a wide variety of astronomical software
used at Lowell Observatory and elsewhere.
(3327) Campins
1985 PW. Discovered 1985 August 14 by E. Bowell at Anderson Mesa
Named in honor of Humberto Campins, research scientist at the
Planetary Science Institute in Tucson. Well known for his work on the
properties of cometary comae, Campins has helped establish pioneer-
ing techniques to measure the physical properties of cometary nuclei
using simultaneous infrared and visual observations. He has also
undertaken infrared searches for intramercurial bodies.
(4551) Cochran
1979 MC. Discovered 1979 June 28 by E. Bowell at Anderson Mesa
Named in honor of William D. Cochran and Anita L. Cochran,
husband and wife astronomers at the University of Texas at Austin.
A
p
p
e
n
d
i
x

1
212
Williams broad range of research has concerned planetary (includ-
ing cometary) atmospheres, Raman scattering, stellar radial velocity
variations and motions in stellar chromospheres and photospheres.
Using a spectroscopic radial velocity meter, he is currently surveying
several dozen stars to search for reex motions (as small as about 2
m/s) that would indicate the presence of planetary companions. Anita
is a specialist in the chemistry of cometary comae and in particular
how the chemistry changes with changing heliocentric distance. An
assiduous observer, she has used spatially resolved spectra to help
transform the photometry of comets into a quantitative discipline.
She has also developed sophisticated cometary models to understand
how the observed atoms, molecules, and radicals are related to the
larger parent molecules present in cometary nuclei. Anita is cur-
rently a team member of the Imaging Science Subsystem of the
Comet Rendezvous Asteroid Flyby mission.
(3531) Cruikshank
1981 FB. Discovered 1981 March 30 by E. Bowell at Anderson Mesa
Named in honor of Dale P. Cruikshank, planetary scientist at the
University of Hawaii, Honolulu. Cruikshank is well known for his
observational work on solar system small bodies, including Trojan
asteroids, comets and Pluto. He is especially known for studies of
outer-planet satellites, including Triton, Iapetus and Io, through both
telescopic and Voyager spacecraft observations. He has been active in
developing instrumentation and facilities at Mauna Kea Observatory
and has promoted historical studies of planetary science. Through
several extended working visits to the Soviet Union and other pro-
jects, Cruikshank has also been a leader in furthering international
scientic relations.
(3638) Davis
1984 WX. Discovered 1984 November 20 by E. Bowell at Anderson
Mesa
Named in honor of Donald R. Davis, senior scientist at the Planetary
Science Institute in Tucson. Davis has made fundamental theoretical
and experimental contributions to research on the collisional evolu-
tion of minor planets. With colleagues, he was the rst to propose the
gravitationally bound rubble pile model for large minor planets.
D
r
a
m
a
t
i
s

p
e
r
s
o
n
a
e
213
Another of his research interests is infrared searching for intramer-
curial bodies.
(9064) Johndavies
1993 BH8. Discovered 1993 January 21 by the Spacewatch at Kitt Peak
John K. Davies (1955 ) of the Joint Astronomy Centre was instrumen-
tal in the successful discovery and follow-up of (3200) Phaethon and
several comets with the IRAS satellite in 1983. He has also carried out
studies of the infrared nature of distant minor planets and authored a
number of popular books and articles.
(6115) Martinduncan
1984 SR2. Discovered 1984 September 25 by B. A. Skiff at Anderson
Mesa
Named in honor of Martin J. Duncan (1950 ) of Queens University,
Kingston, Ontario. Duncan has made several important contributions
to the understanding of the origin and dynamical evolution of small
bodies in the solar system, particularly comets and the likelihood that
they originated in the Kuiper Belt. He has been involved in the devel-
opment of two important numerical algorithms that have led to
orbital integrations of unprecedented duration.
(3487) Edgeworth
1978 UF. Discovered 1978 October 28 by H. L. Giclas at Anderson Mesa
Named in memory of Kenneth Essex Edgeworth (18801972), Irish
engineer, economist, military man and independent theoretical
astronomer, who reasoned that the solar system did not end with
Neptune. As early as 1943 he pointed out the likely existence of a
reservoir of potential comets near the invariable plane. This preceded
the discovery of 1992 QB
1
by almost half a century.
(2664) Everhart
1934 RR. Discovered 1934 September 7 by K. Reinmuth at Heidelberg
Named in honor of Edgar Everhart (19201990), since 1969 in the
physicsastronomy department at the University of Denver and direc-
A
p
p
e
n
d
i
x

1
214
tor of the Chamberlain Observatory. After an impressive career
working on atomic cross-sections, he has made equally fundamental
contributions to our knowledge of the distribution of comets and the
evolution of cometary orbits, including the development of an
efcient integration technique for the purpose. Visual discoverer of
comets 1964 IX and 1966 IV, he has more recently designed and con-
structed a measuring engine and used it in a highly successful
program of photographic astrometry of comets.
(3248) Farinella
1982 FK. Discovered 1982 March 21 by E. Bowell at Anderson
Mesa
Named in honor of Paolo Farinella, planetary scientist at the
University of Pisa, whose research has included studies of the origin
of the solar system and the dynamics of planetary satellites and ring
systems. Farinellas work on minor planets has concerned the colli-
sional evolution of the belt and the formation of families, both from a
theoretical and an experimental point of view.
(5996) Julioangel
1983 NR. Discovered 1983 July 11 by E. Bowell at Anderson Mesa
Named in honor of Julio Angel Fernandez (1946 ) of the
Universidad de la Republica, Montevideo. Fernandez is a noted
dynamicist who has worked on the evolution of comet orbits and
planetesimal scattering in the outer solar system, including the for-
mation of the Oort Cloud. His work has led to some of the rst clear
indications for the existence of the trans-Neptunian belt. Since 1985
he has contributed to the reestablishment of Uruguayan astronomy
by educating a vigorous group of young planetary scientists and
dynamicists.
(4985) Fitzsimmons
1979 QK
4
. Discovered 1979 August 23 by C.-I. Lagerkvist at La
Silla
Named after Alan Fitzsimmons, who works on the relationships
between minor planets and comets and has collaborated with the
D
r
a
m
a
t
i
s

p
e
r
s
o
n
a
e
215
discoverer for several years. His enthusiasm and good spirit when
observing on La Palma has always been very much appreciated by the
discoverer.
(1777) Gehrels
4007 P-L. Discovered 1960 September 24 by C. J. van Houten and I. van
Houten-Groeneveld at Palomar
Named in honor of Tom Gehrels (1925 ), staff member of the Lunar
and Planetary Laboratory at Tucson. Dr Gehrels is well known for his
photometric and polarimetric observations of minor planets and the
Moon.
(7728) Giblin
1977 AW
2
. Discovered 1977 January 12 by E. Bowell at Palomar
Named in honor of Ian Giblin (1969 ), a British physicist who has per-
formed a number of laboratory experiments to simulate hyperveloc-
ity impacts among minor planets. Giblin has developed new data
analysis tools to study their outcome and to draw conclusions regard-
ing the corresponding actual events.
(7638) Gladman
1984 UX. Discovered 1984 October 26 by E. Bowell at Anderson Mesa
Named in honor of Brett Gladman (1966 ), a Canadian astronomer
and dynamicist who has made important contributions to modelling
the dynamical evolution of near-Earth objects and the transport of
meteorites, including those from the Moon and Mars. Gladman has
also carried out observational surveys of trans-Neptunian objects
and in 1997 was codiscoverer of the two irregular satellites of
Uranus.
(2068) Dangreen
1948 AD. Discovered 1948 January 8 by M. Laugier at Nice
Named in honor of Daniel W. E.Green (1958 ), student aide at the
Smithsonian Astrophysical Observatory during 1978 JuneAugust, in
appreciation of his invaluable assistance during the transition of the
Minor Planet Center from Cincinnati to Cambridge [Massachussetts].
A
p
p
e
n
d
i
x

1
216
(3676) Hahn
1984 GA. Discovered 1984 April 3 by E. Bowell at Anderson Mesa
Named in honor of Gerhard Hahn, a planetary astronomer at Uppsala
Observatory and a member of the research group studying minor
planets and comets. Hahn has undertaken extensive photometry and
astrometry of minor planets and has been studying the long-term
orbital evolution and physical properties of these objects.
(3267) Glo
1981 AA. Discovered 1981 January 3 by E. Bowell at Anderson Mesa
Named in honor of Eleanor F. (Glo) Helin, planetary scientist at the
Jet Propulsion Laboratory, in appreciation of her extraordinary con-
tributions to the discovery of near-Earth minor planets. Her nding of
1976 AA (=2062 Aten) heralded the recognition of a new class of
planet-crossers, and her initiation of the Palomar planet-crossing
asteroid survey has resulted in increased worldwide interest in the
observation of minor planets. Helins education and experience as a
geologist and in the analysis of meteorites has provided a unique
background for her interest in asteroids and comets.
(3099) Hergenrother
1940 GF. Discovered 1940 April 3 by Y. Visl at Turku
Named in honor of Carl William Hergenrother (1973 ) of the Bigelow
Sky Survey. This photographic survey has been very successful in dis-
covering new high-inclination minor planets.
(4205) David Hughes
1985 YP. Discovered 1985 December 18 by E. Bowell at Anderson Mesa
Named in honor of David W. Hughes, reader in physics at Shefeld
University, where he teaches courses on all aspects of astronomy. His
research area concerns small solar-system bodies, particularly the
relationship between comets and meteors. He has served astronomy
in Britain as a vice president of both the Royal Astronomical Society
and the British Astronomical Association. Hughes is a prolic
reviewer of astronomy books and writes regularly on current issues
in astronomy for Nature.
D
r
a
m
a
t
i
s

p
e
r
s
o
n
a
e
217
(6434) Jewitt
1981 OH. Discovered 1981 July 26 by E. Bowell at Anderson Mesa
Named in honor of David Jewitt (1958 ) of the Institute for
Astronomy, University of Hawaii. The consummate astronomer,
Jewitt has been devoted to astronomy from a very early age. He has
made several important contributions to planetary astronomy, start-
ing with his discovery of the Jovian satellite Adrastea from Voyager
data in 1979. He was co-recoverer of comet 1P/Halley in 1982. Jewitt is
perhaps best known for co-discovering the rst body in the Kuiper
Belt in 1992 (see [minor] planet (1776)), thus proving that accretion
occurs beyond the planetary region. Jewitts main area of research is
comets, but his wide-ranging interests have also produced work on
planetary rings, minor planets, Pluto and circumstellar discs.
(1776) Kuiper
2520 P-L. Discovered 1960 September 24 by C. J. van Houten and I. van
Houten-Groeneveld at Palomar
Named in honor of G. P. Kuiper (19051973), former Director of the
Lunar and Planetary Laboratory at Tucson, and former Director of
the Yerkes Observatory. Dr Kuiper was a well-known authority on the
solar system and initiated both the McDonald Survey and the
PalomarLeiden Survey of minor planets.
(6909) Levison
1991 BY
2
. Discovered 1991 January 19 by C. S. Shoemaker and E. M.
Shoemaker at Palomar
Named in honor of Harold Levison (1959 ) of the Boulder, Colorado,
ofce of the Southwest Research Institute. Since 1988 Levison has, in
collaboration with Martin Duncan, virtually revolutionised our view
of the dynamics of short-period comets. The work uses sophisticated
numerical models of test particles perturbed by the planets, and it has
revealed important details about the Kuiper Belt, Centaurs,
PlutoCharon and short-period comets. Levison has also contributed
revealing insights into the dynamics of perturbers in the Beta
Pictoris system and was a leading member of the team that used
Hubble Space Telescope to discover possible observational evidence
for small comets in the trans-Neptunian region.
A
p
p
e
n
d
i
x

1
218
(5430) Luu
1988 JA
1
. Discovered 1988 May 12 by C. S. Shoemaker and E. M.
Shoemaker at Palomar
Named in honor of Jane X. Luu (1963 ) for her research on the small
bodies of the solar system. Luu is best known for her work with David
Jewitt in discovering the rst and subsequent members of the Kuiper
Belt, as well as in following up with physical studies of those bodies.
She has also contributed the most stringent upper limits on the exis-
tence of dusty comae around minor planets that might be dormant or
extinct comets.
(6698) Malhotra
1987 SL
1
. Discovered 1987 September 21 by E. Bowell at Anderson Mesa
Named in honor of Renu Malhotra (1961 ), accomplished dynamicist
and celestial mechanician at the Lunar and Planetary Institute in
Houston. Born and raised in India, she has made major contributions
to our understanding of how resonances affect satellite systems, the
asteroid belt, and particularly Pluto. Malhotra was awarded the
Harold C. Urey Prize by the Division for Planetary Sciences of the
American Astronomical Society in 1997. Her talents and good spirits
are much enjoyed by her colleagues.
(1877) Marsden
1971 FC. Discovered 1971 March 24 by C. J. van Houten and I. van
Houten-Groeneveld at Palomar
Named in honor of Brian G. Marsden (1937 ), Smithsonian
Astrophysical Observatory, in recognition of his numerous contribu-
tions in the eld of orbit calculations for comets and minor planets,
his improved versions of the Catalogue of Cometary Orbits, and his
activities in the Central Bureau and in Commission 20 of the IAU.
(4367) Meech
1981 EE
43
. Discovered 1981 March 2 by S. J. Bus at Siding Spring
Named in honor of Karen J. Meech (1959 ) of the Institute for Astron-
omy of the University of Hawaii for her pioneering studies of comets
very far from the Sun. Her work following new and long-period comets
D
r
a
m
a
t
i
s

p
e
r
s
o
n
a
e
219
to great distance has been a major factor in changing our ideas about
water as the predominant driver of cometary activity in most comets.
Meechs studies of (2060) = 95P/Chiron have likewise been critical in
changing our understanding of the nature of the cometary coma.
(7639) Offutt
1985 DC
1
. Discovered 1985 February 21 at the Oak Ridge Observatory at
Harvard
Named in honor of Warren Offutt (1928 ), on the occasion of his 70th
birthday, 1998 February 13. After a career as an engineering executive,
he turned in his retirement to the astronomical applications of CCDs,
considering in particular the contributions that can be made by
amateur astronomers. At his observatory in New Mexico he has made
key observations of several of the objects discovered in the Kuiper
Belt in recent years, as well as of other comets and minor planets as
faint as 22nd magnitude. His follow-up of S/1997 U 2, one of the two
recently discovered satellites of Uranus, played a crucial role in the
establishment of its orbit.
(5040) Rabinowitz
1972 RF. Discovered 1972 September 15 by T. Gehrels at Palomar
Named in honor of the US astronomer David Rabinowitz and his work
in the Spacewatch program.
(3594) Scotti
1983 CN. Discovered 1983 February 11 by E. Bowell at Anderson Mesa
Named in honor of James V. Scotti (1960 ) of the University of
Arizona, Tucson. Scotti works with the SPACEWATCH Telescope,
which is the 0.9 m reector of the Steward Observatory on Kitt Peak.
He has developed most of the systems software and has carried out
nal checks and data reduction for the CCD scanning observations of
comets and minor planets.
(4446) Carolyn
1985 TT. Discovered 1985 October 15 by E. Bowell at Anderson Mesa
Named inhonor of CarolynSpellmannShoemaker, comet and aster-
oid discoverer. Shoemaker begansearching for asteroids in1980, using
A
p
p
e
n
d
i
x

1
220
plates takenat the UKSchmidt Telescope at Siding Spring. She helped
develop a newphotographic survey programusing the 0.46 mSchmidt
camera at Palomar Mountainand a newly designed stereomicroscope,
whichgreatly increased the efciency of lmscanning. In1983
Shoemaker found her rst near-Earthasteroid, the Amor object (3199)
Nefertiti, and later that year she found her rst comet, 1983p. By
February 1991 she had discovered 22 comets, at a rate of about one per
100 hours of scanning, and for discoveries recognised inthe names of
the comets she thus surpassed the tally of W. R. Brooks and moved into
the all-time second place behind J.-L. Pons. Shoemaker already holds
the record for nding newperiodic comets: nine by early 1991.
(10234) Sixtygarden
1997 YB
8
. Discovered 1997 December 27 by J. Tich and M. Tich at Klet
The street address of the Harvard-Smithsonian Center for
Astrophysics is 60 Garden Street. Observers of minor planets and
comets know it as the seat of the Minor Planet Center (see planet
(4999)) and the Central Bureau for Astronomical Telegrams, which
communicate fast-breaking news of astronomical discoveries to the
international community.
(7554) Johnspencer
1981 GQ. Discovered 1981 April 5 by E. Bowell at Anderson Mesa
Named in honor of John R. Spencer (1957 ), an astronomer at Lowell
Observatory, for his pioneering interdisciplinary work in planetary
science. Spencers research includes detailed and insightful probing
into the nature and character of the Galilean satellites. In particular,
he has led the eld in using high-resolution, ground-based imaging of
Io to provide an excellent time history of volcanism, important for
bridging the gaps between spacecraft encounters. In his studies,
Spencer applies a keen intuitive sense of the natural world and leaves
us with a better appreciation and understanding of our solar system.
(2309) Mr Spock
1971 QX
1
. Discovered 1971 August 16 by J. Gibson at El Leoncito
Named for the ginger short-haired tabby cat (1967 ) who selected the
discoverer and his soon-to-be wife at a cat show in California and
D
r
a
m
a
t
i
s

p
e
r
s
o
n
a
e
221
accompanied them to Connecticut, South Africa and Argentina. At El
Leoncito he provided endless hours of amusement, brought home his
trophies, dead or alive, and was a gure of interest to everyone who
knew him. He was named after the character in the television
program Star Trek who was also imperturbable, logical, intelligent
and had pointed ears.
(6373) Stern
1986 EZ. Discovered 1986 March 5 by E. Bowell at Anderson Mesa
Named in honor of S. Alan Stern (1957 ) of the Boulder, Colorado,
ofce of Southwest Research Institute. Sterns research has focused
on both observational and theoretical studies of the satellites of the
outer planets, Pluto, comets, the Oort Cloud and the Kuiper Belt (see,
respectively, planets (1691) and (1776)). He is also active in instrument
development, with a strong concentration in ultraviolet and imaging
technologies. He has participated in ten planetary sounding-rocket
missions, two Space Shuttle mid-deck experiments and a Shuttle-
deployable satellite. He was chair of NASAs Outer Planets Science
Working Group during 19911994.
(4438) Sykes
1983 WR. Discovered 1983 November 29 by E. Bowell at Anderson Mesa
Named in honor of Mark V. Sykes, planetary scientist at the Steward
Observatory of the University of Arizona, Tucson. Sykes was the rst
to suggest that the dust bands discovered in data from the Infrared
Astronomical Satellite (IRAS) were due to the catastrophic disrup-
tions of small asteroids and comets. He has also discovered several
additional dust bands, a second type of dust trail, and identied
parent comets responsible for some of the IRAS dust trails.
(3255) Tholen
1980 RA. Discovered 1980 September 2 by E. Bowell at Anderson Mesa
Named in honor of David J. Tholen (1955 ), planetary scientist at the
Institute for Astronomy of the University of Hawaii. Tholens work on
the eight-colour survey of minor planets led him to devise an
improved taxonomy of minor planets. He has considered the physical
properties of minor planets, satellites and comets in terms of compo-
A
p
p
e
n
d
i
x

1
222
sition and evolution, and he was among the rst to observe events in
the series of occultations and transits now occurring between Pluto
and its satellite Charon.
(1604) Tombaugh
1931 FH. Discovered 1931 March 24 by C. W. Tombaugh at Flagstaff
Named by the Lowell Observatory after Clyde W. Tombaugh
(19061997), the discoverer of Pluto, on the occasion of a symposium
on Pluto, held on the ftieth anniversary of its discovery, 1980
February 18. Tombaugh marked, during the course of his blink exami-
nation, over 4000 minor planets on plates obtained with the 0.33 m
photographic telescope during the trans-Saturnian search program at
the Lowell Observatory
(3634) Iwan
1980 FV. Discovered 1980 March 16 by C.-I. Lagerkvist at La Silla
Named in honor of Iwan P. Williams, of Queen Mary College, London,
in recognition of his well-known work on meteor streams and interest
in comets and minor planets. The discoverer appreciates their long
and fruitful collaboration.
(1940) Whipple
1975 CA. Discovered 1975 February 2 at the Harvard College
Observatory at Harvard
Named in honor of Fred L.Whipple (1906 ), Harvard astronomer
since 1931, professor since 1950 and director of the Smithsonian
Astrophysical Observatory from 1955 to 1973. His countless contribu-
tions to our knowledge of the smaller bodies of the solar system
include his icy-conglomerate model for cometary nuclei, and the
development of modern techniques for the photographic observations
of meteors. He has served as president of IAU Commissions 6, 15, and
22, and is now active on the NASA panel of space missions to comets
and minor planets.
D
r
a
m
a
t
i
s

p
e
r
s
o
n
a
e
223
Guidelines for
minor planet
names
Introduction
All minor planet names are subject to the approval of the Committee
for Small Body Nomenclature (CSBN), a working group of IAU
Division III, in consultation with the IAUs Minor Planet Center
(MPC). The guidelines are based on resolutions that have been passed
at various times by IAU Commission 20, to which the CSBN and its
predecessor committees reported until 1997. The CSBN advises the
MPC on how to interpret these resolutions in individual cases. We
provide here the relevant resolutions from the Transactions of the
IAU. Note that the reviewing structure and schedule have changed
over the period covered by these resolutions. All reviews are now
carried out by the CSBN, which was created in 1994.
Montreal, 1979 FromProceedings of the General Assembly XVIIA
Commission 20 afrms the tradition that the discoverer of a numbered
minor planet be permitted to propose a name for the object. A proposal
would normally be accepted, provided that it is consistent with the broad
policy of the Working Group on Planetary System Nomenclature. The
name would become ofcial following the publication in the MPCs of a
brief citation explaining its signicance. The Commission denes the dis-
covery as the earliest apparition at which an orbit useful in the establish-
ment of identications was calculated; in the case of double designations
during the same apparition, priority will be given in order of announce-
ment of discovery, unless the double designation follows from an orbit
computation using the observations made according to the second
announcement. Further, the Commission proposes that, if the discoverer is
deceased, or if a minor planet remains unnamed ten years after it has been
numbered, a name could appropriately be suggested by identiers of the
A
P
P
E
N
D
I
X

2
225
various apparitions of the object, by discoverers at apparitions other than
the ofcial one, by those whose observations contributed extensively to the
orbit determination, or by representatives of the observatory at which the
ofcial discovery was made. In such a case, the selection of a name shall be
judged by a committee of three, consisting normally of the President and
Vice President of the Commission 20 and the Director of the Minor Planet
Center, and the nal decision shall be made not less than six months follow-
ing the announcement of the numbering of the minor planet.
Patras, 1982 From Transactions of the IAUXVIIIB
All names proposed for minor planets will be reviewed for suitability, even
when names are proposed by discoverers. The review will be done as indi-
cated in the 1979 Commission 20 resolution, except that in the case of a
name proposed by the discoverer, the six-month waiting period for a newly
numbered object can be reduced to two months. Names shall be limited to a
maximum length of sixteen characters, including spaces and hyphens.
[The review committee is composed of the President and the Vice President
of the Commission and the Director of the Minor Planet Center.]
The last sentence, in brackets, was part of the original resolution
but was later changed before adoption.
Delhi, 1985 From Transactions of the IAU XIXB
Names proposed for minor planets will not be accepted if, in the opinion of
the Minor Planet Names Committee, they are too nearly similar to those of
other minor or major planets or natural satellites, or are in questionable
taste. Names should be pronounceable, preferably expressible as a single
word, and no more than sixteen characters long. Names glorifying individ-
uals or events principally known for their political or military activities or
implications are considered unsuitable unless at least one hundred years
have elapsed since the individuals died or the events concerned took place.
Objects involved with the Jovian triangular libration points should be
named in accordance with the tradition of honoring heroes of the Trojan
War. In a disputed case, the proposer may appeal the committees decision
at a general meeting of Commission 20, provided that due written notice is
given to the President of the Commission.
Summary of guidelines
Discoverers have the privilege to propose names for ten years
after numbering. Beyond that point, others may propose names.
Names must be pronounceable, preferably expressible as a
single word, and no more than 16 characters in length.
A
p
p
e
n
d
i
x

2
226
Individuals or events principally known for political or military
activities are unsuitable until 100 years after the death of the
individual or the occurrence of the event.
Names of pet animals are discouraged.
Minor planets in certain dynamical groups should be named
within more restrictive guidelines. For example,
Trojan asteroids (those that librate in 1:1 resonance with
Jupiter) are named for heroes of the Trojan War (Greeks at
L4 and Trojans at L5).
Trans-Jovian planets crossing or approaching the orbit of a
giant planet but not in a stabilising resonance are named
after centaurs.
Objects crossing or approaching the orbit of Neptune and in
stabilising resonances other than 1:1 are given mythological
names associated with the underworld. (Planned guideline
to follow the example of Pluto.)
Objects sufciently outside Neptunes orbit such that
orbital stability is reasonably assured for a substantial frac-
tion of the lifetime of the solar system are given mythologi-
cal names associated with creation. (Planned guideline
none yet named.)
Objects that approach or cross Earths orbit are given
mythological names.
The SBNC may choose to act on its own in naming a minor
planet and has traditionally done so in each case in which the
number is an integral number of thousands.
This material kindly supplied by Mike AHearn on behalf of the IAU
Committee for Small Body Nomenclature.
G
u
i
d
e
l
i
n
e
s

f
o
r

m
i
n
o
r

p
l
a
n
e
t

n
a
m
e
s
227
55 Cancri, dust disc 177
200 inch telescope, see Hale Telescope
1977 UB see Chiron
1992 AD see Pholus
1992 QB
1
colour of 117
discovery of 65, 66, 67
naming of 2023
orbit of 68, 69, 85
1993 FW
colour of 117
discovery of 70
naming of 2023
1993 HA
2
, see 7066 Nessus
1993 RO
discovery of 71
orbit of 95, 96, 97
1993 RP
discovery of 71
orbit of 95, 97, 141
1993 SB
discovery of 723
naming of 203
orbit of 95, 97, 98, 100
1993 SC
discovery of 73
naming of 203
orbit of 95, 97
physical observations of 127, 130, 131,
132, 144
1995 GO see Asbolus
1995 SM
55
, discovery of 92
1996 RR
20
, discovery of 80
1996 TL
66
discovery of 85
orbit of 86, 87, 88, 111
physical observations 1323, 136, 144
1996 TO
66
physical observations 129, 133, 135
size 145
1996 TP
66
, physical observations 135
1996 TQ
66
, physical observations 135
1996 TR
66
, orbit of 98
1997 CU
26
see 10199 Chariklo
1997 GA
45
, discovery of 82
1997 RL
13
, discovery of 81
1997 RT
5
, discovery of 80
1999 DG
8
, discovery of 82, 83
2000 AC
255
, discovery of 82
2000 AF
255
, discovery of 82
7066 Nessus 35
8504 Asbolus 35, 119
944 Hidalgo 28
10199 Chariklo 35
albedo 143
AHearn, Mike 202, 204, 207, 211
albedo 141
of asteroids, etc. 142, 143, 145
angular momentum, conservation of 12,
108
aperture correction 124
aphelion 16
apparent motion vector 61, 80, 83
Asbolus 35, 119
ascending node 105
asteroid belt 4, 118, 1812
asteroid naming 25, 27
astrometry 39
astronomical Unit (AU) 16
Bailey, Mark 38, 192, 201, 209, 211
Barucci, Antonietta 117, 128, 212
Bernstein, Gary 61, 82, 164
229
Index
Page numbers in italics refer to gures.
Beta Pictoris disc 1689, 170
Big Throughput Camera, see BTC
bimodality of Kuiper Belt object colours
120, 121, 1356, 157
blink comparator 5, 23, 51
blinking CCD images 56, 57, 64, 72, 80
blinking photographs 6, 9, 22, 51, 145
Brown, Mike 82, 135
Brown, Robert (Bob) 131, 177
BTC 83
CanadaFranceHawaii Telescope, see
CFHT
CCD
data processing 55, 61, 80, 123, 1467
operation of 524, 89
Centaurs 356, 92
as inactive comets 37
mythology of 27, 32
orbital lifetimes of 38
orbits of 37
Central Bureau for Astronomical
Telegrams 22
CFHT 58, 79, 82
Charge Couple Device, see CCD
Chariklo, see 10199 Chariklo
Charon (moon of Pluto) 10, 161,
206
Chiron
albedo of 143
as a comet 289, 30, 368, 115, 202
discovery of 24, 25
naming of 27, 202
orbit of 26, 28
physical observations 28, 30, 34, 35
Classical Kuiper Belt, naming of 85
Cloudcroft observatory 76, 77, 86
Cochran, Anita 60, 76, 145, 212
Cochran, Bill 60, 174, 212
collisions
between asteroids 151, 153
computer models of 155
experiments 1536
in Kuiper Belt 119, 1512, 159, 161
colours of Kuiper Belt objects 11720, 121,
130
coma 17
comet belt, hypothetical 2, 1415, 41, 42,
46
comet
naming 25
nucleus 17
tails 18
cometary outbursts 115, 119, 126, 129,
136
comets
orbits of 16, 17, 19
physical properties 16, 17
commensurability 94, 95
computer simulations 40, 445, 1004, 109,
179
computerised searches 59, 61, 86, 89, 91, 92,
183
coronagraph 169, 176, 177
cosmic rays 117, 119
effects on CCDs 146
Cruikshank, Dale 36, 131, 195, 196, 213
Cubewano 85
cumulative luminosity function 145, 149,
150
Davies, John 136, 143, 214
Davis, Don 151, 152, 153, 154, 161, 213
denition of a planet 207
Delahodde, Catherine (Cathy) 125, 128
descending node 105
drift scanning (of CCD) 89
Duncan, Martin 43, 44, 1003, 111, 179, 214
Kuiper Belt object searches 58, 60, 146
on names 191, 199, 204
dust detection experiments 1579
dust discs around other stars 16878
dust in Kuiper Belt 1579
eccentricity 93
ecliptic 8
Edgeworth, Kenneth Essex 1, 2, 10, 15, 42,
85, 192, 1969, 214
EdgeworthKuiper objects, 199
see also Kuiper Belt objects
Epsilon Eridani 173, 174
Everhart, Edgar 20, 39, 40, 42, 45, 214
Farinella, Paulo 151, 152, 161, 215
Felix 203
Fernandez, Julio 412, 111, 215
on names 197, 208
Fitzsimmons, Alan 724, 80, 127, 215
on names 203, 206
at elding, see CCD, data processing
Gareld 203
Gehrels, Tom 25, 30, 878, 91, 195, 216
Gladman, Brett 76, 79, 164, 216
Greaves, Jane 1723
Green, Dan 24, 193, 198, 209, 216
Green, Simon 117, 136
Hainaut, Olivier 128, 130
Hale Telescope 79, 80
I
n
d
e
x
230
Halleys comet 25, 28
recovery of 48
Helin, Eleanor Glo 31, 81, 217
HR4796A 175, 176, 177
HST
Kuiper Belt search 1469
observations of dust discs 170, 176
physical observations of Centaurs, 119
physical observations of Kuiper Belt
objects 126, 135
Hubble Space Telescope, see HST
Hut, Piet 60
IAU
circulars 23, 66, 69, 71, 83
telegrams 23
see also International Astronomical
Union
ice dwarfs 162
inclination 19
infrared cameras 143, 175
infrared emission
from dust around stars 167, 168, 175
from Kuiper Belt objects and centaurs
1424,
from solar system dust 157
infrared spectra 345, 36, 131, 132, 135
INT 72, 73, 97
integration of orbital elements, 99, 111
International Astronomical Union 22,
206
invariable plane 8
irradiation mantle 118
IRTF telescope 143, 177
Isaac Newton Telescope, see INT
ISO observations 143
James Clerk Maxwell Telescope see JCMT
JCMT 171, 177
Jewitt, Dave
early career 479, 57
early searches 50, 51, 546
on future research 181
on names 193, 199, 208, 218
physical observations of KBOs 117, 130,
1323, 136, 143
searches from Mauna Kea 58, 62, 645, 85,
87
Joss, Paul 20
Jupiter, gravitational inuence of 39, 45,
108
Jupiter family comets 19, 45, 113
Karla 202, 203
Keck Observatory 97, 1302, 1356
Kowal, Charles 22, 23, 267, 29, 145
Kuiper Belt
mass of 149, 159
naming of 46, 68, 1919
structure of 1036, 104, 11213
thickness of 151
Kuiper Belt objects
colours 11721, 1356
growth of 159
lightcurves of 1279
numbers of 149
sizes of 1445
variability of 12730, 133, 136
Kuiper Belt telescope 183
Kuiper, Gerard 3, 10, 15, 17, 57, 85, 1926,
218
Lagrangian points 96
Leonard, F.C. 200
Leuschner, Otto Armin 199200
Levison, Hal 58, 60, 103, 111, 146, 218
on names 201, 204
libration 95
lightcurves 33, 127, 128, 129
long-period comets, orbits 16
Lowell Observatory 5, 6, 205
Lowell, Percival 5
Luu, Jane
physical observations of Kuiper Belt
objects 117, 130, 133, 136
see also Jewitt, Dave 49, 50, 85, 208, 219
magnitude system 33
Malholtra, Renu 10910, 219
Marsden, Brian 15, 23, 24, 26, 66, 68, 70, 76,
96, 141, 200, 219
on Plutos status 2034
MASCOT camera 54
Mauna Kea, as an observatory site 578, 62,
116
mean motion resonance 946, 98, 106, 109,
111, 112, 113
Meech, Karen 29, 128, 219
Monte-Carlo methods 40
Monthly Notices of the RAS 3, 42, 193
migration of planets 10710
millimetre wave observations 1713
Minor Planet Center 24, 201, 2256
minor planet 10000 204
minor planet names 25, 201, 2256
minor planets
colours see Kuiper Belt objects, colours
designations 25, 66fn
faintest designated 82
most distant 82
I
n
d
e
x
231
Neptune
discovery of 4
gravitational inuence of 42, 45, 945,
103, 105, 110
orbital migration of 1078
Nessus 35
observational biases 1401
observed arc 26
Offutt, Warren 76, 77, 78, 86, 220
Oort cloud, 17, 18, 41
naming of 17, 199
Oort, Jan 16
orbital elements 74, 99, 103
pencil-beam survey 79
perihelion 16
Pholus
albedo 1423
discovery of 31
naming of 32, 202
observational properties 32, 345, 367,
135, 136
orbit of 32
photometry 33
blind test 125
infrared 1356, 1434
techniques 1225, 127
Pioneer 10 spacecraft 158
pixel 53
planet formation 2, 1314
Planet X 5, 6, 8, 10, 205
planetesimal
as protocomet 16
formation 1314
Plutinos
fraction of 140
naming of 84
number of 149
orbital stability of 1036
Pluto
discovery of 9
formation of 161
gravitational inuence on comets 3,
17
missions 1848
numbering of 2039
orbit of 10, 935
planetary status of 2039
rotation period of 10
satellite of, see Charon
size of 9, 10, 15, 205
surface properties 132, 142
PlutoKuiper Express, see Pluto, missions
power law index 139
PoyntingRobertson drag 157, 177
precession 103
precovery (of asteroids) 26
Rabinowitz, David 31, 32, 81, 91, 220
recovery (of asteroid) 26
reex motion 50, 146
resonance, see mean motion resonance
resonance sweeping 110, 118
resurfacing of KBOs 119, 12930
Romanishin, Bill 120, 125, 127, 136
rotation periods of Kuiper Belt objects
1279
Ryan, Eileen 1524, 208
scattered disc objects 87, 88, 92, 111, 112,
113
number of 149
Scotti, James (Jim) 90, 91, 92, 205, 220
SCUBA 1725
searches for Kuiper Belt 50, 51, 5861, 645
secular resonance 103, 105, 106
seeing 56
effects of 57, 60, 122, 123
selection effects 13941
Shoemaker, Carolyn 31, 220
short-period comets
capture of 19, 20
excess of 19, 39, 41
lifetimes of 19
orbits of 17, 19
origin of 19, 20, 446, 112
size distribution 79, 139
Smiley 2023
Smithsonian Astrophysical Observatory
23
software aperture 123, 124
solar system, formation of 1114
Spacewatch telescope 30, 35, 87, 8891, 90
spectroscopy 34
infrared 346, 1313, 134
visible 133, 134
Stern, Alan 146, 159, 164, 222
on names 191, 201, 205, 207
TaiwanAmerican Occultation Survey
1801
Tegler, Steve 1201, 125, 127, 136
thermal emission, see infrared emission
Tholen, David 289, 32, 76, 97, 122, 136,
222
tholins 35
Tombaugh, Clyde 6, 7, 89, 145, 208, 223
Tremaine, Scott 44, 191
Trojan asteroids 96
I
n
d
e
x
232
Trujillo, Chad 85, 86, 87, 149, 164
two-colour diagram 33, 121, 130
UBVRI lters 33
UH 2.24 m telescope 58, 59, 62, 63, 85, 97, 136
UKIRT 34, 58, 136, 143
University of Hawaii see UH
Uranus
deviations from predicted position 4, 10
discovery of 4
variability of KBOs 127, 129
Vega excess stars 16872
Voyager spacecraft 158
Whipple, Fred 14, 15, 17, 199, 200,
223
Williams, Iwan 72, 100, 127, 203, 223
water ice 356, 133, 135, 136
zodiacal dust 144, 157, 158
I
n
d
e
x
233