The EconomistDecember 21st 2019 Holiday specials 37I
t is hardlysurprising that in the 17th and 18th centuries scien-
tists likened the movements of the solar system to the ticking of
well regulated machinery. The clockwork of orreries, mechanical
models of the solar system, neatly encapsulated the apparently
clockwork nature of the heavens, each planet following its desig-
nated course just as it always had, world without end.
The beginning, though, was much less orderly. Wind the clock
back far enough and the clockwork goes awry. When the solar sys-
tem was in its infancy, it now seems, planets changed their orbits
with feckless abandon, swinging in towards the sun and out again,
sometimes swapping places, possibly leaving the solar system al-
together. These peregrinations seem to explain long-standing
mysteries about why the solar system is the way it is.
The solar system formed from a cloud of gas and dust which
collapsed into the Sun and a disc of dust and gas from which the
planets grew through accretion and collision, with smaller things
merging into bigger ones until there were no mergers left to be
had. The end result was a neat arrangement of four inner planets—
Mercury, Venus, Earth and Mars—and four outer ones—Jupiter,
Saturn, Uranus and Neptune—all in nearly circular but in fact el-
liptical orbits that kept them strictly separated. The inner ones
were made mostly of rock and relatively small: the Earth is the larg-
est. The outer ones were bigger and made mostly of gas. Disc mate-
rial that did not get swept up into planets formed a belt of asteroids
between the inner and outer quartets and a belt of small, icy bodies
beyond them.
In the absence of other planetary systems to study, this orderly
arrangement was considered likely to be typical. Its division was
taken to be the result of a “snowline” in the original disk. Sunwards
of this snowline it was too hot for volatile compounds like water,
methane, carbon dioxide and ammonia to condense, and so the
planets were mostly bare rock and iron, with some atmosphere
added later on. Beyond the snowline, these volatiles condensed
into giant gas-balls.
In the mid-1990s, though, astronomers started to discover
planets around other stars—and they were arranged very differ-
ently. The most easily detected were “hot Jupiters”, gas giants cir-
cling their stars in orbits much tighter than that of Mercury, the
nearest planet to the Sun, and thus far inside any conceivable
snowline. Other systems featured giant planets in curious, elon-
gated orbits inexplicable in terms of a coagulating disk. Either that
model of how the planets formed was particular to the solar sys-
tem—unlikely—or planets could change their orbits.
The idea that planets might migrate was not entirely new. In
1984 Julio Fernández and Wing-Huen Ip, then working at the Max
Planck Institute for Solar System Research in Germany, used a
computer to study the interactions between the outer planets and
smaller leftover bodies, known as planetesimals, which took place
in the solar system’s early days. When a planetesimal passes close
to a planet it gets flung into a new orbit. On average the three outergiants flung planetesimals inwards, towards the sun.
Jupiter, though, threw its prey outwards. These en-
counters changed the orbits of the planets, too—infini-
tesimally, case by case, but cumulatively quite a lot. As
Jupiter flung things out, it moved in towards the sun.
As the others flung things in, they moved out.
In 1993 Renu Malhotra, now at the University of Ari-
zona, suggested that this idea might explain the orbit
of Pluto, a dwarf planet. Pluto has a high eccentricity (it
is considerably farther from the sun at some times
than at others) and a high inclination (the plain of its
orbit sits at a marked angle to the disk in which the oth-
er planets travel). On top of that, Pluto makes exactly
two orbits of the Sun for every three orbits made by
Neptune, an arrangement called a resonance; such res-
onances make it easy for momentum to move from one
body to another, giving bigger planets a lot of influence
over the orbits of smaller ones. Dr Malhotra proposed
that Neptune had pushed Pluto into its odd orbit as it
migrated outwards from the Sun, and speculated that it
might have done the same to other bodies it came
across as it migrated. Within a few years a number of
other “trans-Neptunian objects” (tnos) in 2:3 resonant
orbits obligingly got themselves discovered.
But it was the weird exoplanets that started people
thinking that the solar system might not be what it
used to be. “Something big had happened in those sys-
tems, and that made it easier to think about the pos-
sibility that something like that might happen in our
system,” says Bill Bottke of the Southwest Research In-
stitute (swri) in Boulder, Colorado.
To observe these possible pasts needed something
beyond telescopes: computer simulations. In the late
1990s Dr Bottke’s colleague Hal Levison wrote a pro-
gram called swiftwhich could follow the orbits of bo-
dies interacting with each other over billions of years.
Together with his colleagues Kleomenis Tsiganis, Rod-
ney Gomes and Alessandro Morbidelli at the Universi-
ty of Côte d’Azur in Nice, Dr Levison used this software
to model the evolution of the planets’ early orbits.
Their results, published in 2005, dramatically illus-
trated a point made by Dr Malhotra: interactions be-
tween the migrating planets might matter a lot. As Ju-
piter moved inwards and Saturn was pushed out, they
passed through a 1:2 orbital resonance which kicked off
all manner of mayhem. Uranus and Neptune were sud-
denly flung outwards into far more distant orbits like
baubles in a storm; in about 50% of the program’s runs
they swapped places, too.
The Nice model, as it became known, explained theHow the planets
got their spots
It’s a detective story—and a violent family
history. How did the planets of the solar
system end up where they are?The solar systemUranusJupiter Saturn
Neptune Pluto Kuiperbelt
Sun 10 20 30 40 50EarthAU Planet size not to scaleThe solar system today1