Planets in history

Before the curses of light pollution and smog, people were more familiar with the night sky than they tend to be today. Planets in the sky were recognized as special by ancient cultures, because they are ‘wandering stars’ that migrate against the background of the ‘fi xed’ stars. Five planets have been known since antiquity:Mercury, Venus, Mars, Jupiter, and Saturn – which are the only ones bright enough to come to the attention of the unaided eye.Of course, the Sun and Moon were obvious too, but the ‘planets’appear as wandering points of light, whereas the Sun and Moon show disks and tended to be regarded differently. Throughout most of humankind’s existence, the Earth was imagined to be the centre of creation, unrelated to objects in the sky, so it was not thought of as a planet.

The intellectual leaps that recognized that the Earth is a ball of rock going round the Sun, that the planets do likewise, and that the Earth is just one of their number were a long time coming. The process was slow, and there were many false dawns. During the 5th century BC, the ancient Greek philosopher Anaxagoras correctly surmised that the Moon is a spherical body refl ecting the light of the Sun, and he was sent into exile on account of his beliefs. In the succeeding centuries, various Chinese astronomers developed similar ideas, but the idea of the Moon as a globe probably did not embed itself into popular consciousness until its appearance through a telescope became known during the 17th century.

As for the planets, they were generally regarded as points of light going round the Earth, until the counterintuitive ‘heliocentric’ view with the Sun as the centre of motion became accepted. The earliest written suggestions that the Earth goes round the Sun occur in Indian texts dating from the 9th century BC, but despite this and subsequent independent suggestions, notably by Hellenic and Islamic sages and eventually by Nikolas Copernicus in 1543, the concept did not achieve ascendancy until the 18th century. Partly on account of his advocacy of the heliocentric theory, Galileo Galilei (who through his telescope had seen mountains on the Moon, the phases of Venus, and four tiny moons orbiting Jupiter)was held under house arrest from 1633 until his death in 1642.

Simply by revealing the planets as tiny but discernible discs,whereas the stars remained as points of light, use of the telescope from the start of the 17th century onwards marked planets as fundamentally different to stars, and eased the path to regarding them as worlds comparable to our own. Incidentally, we now know that stars are much bigger than planets, but (except for the Sun) they are so very much more distant that only in a few cases can even the most sophisticated of modern telescopes show any surface details (on photographs, bright stars look bigger than faint stars, but that is just an optical effect – the brightness is being smeared out).

Kepler’s laws of planetary motion

The planets slotted into their rightful place in human comprehension thanks to Johannes Kepler’s (1609) realization that the planets (including the Earth) travel round the Sun in paths (orbits) that are ellipses rather than perfect circles, coupled with Isaac Newton’s (1687) insight into gravity that explained this motion. Then their distances and sizes relative to the Earth could begin to be deduced.

An ellipse is what you might think of as an ‘oval’. Mathematically, it is defi ned as a closed curve drawn about two points (the foci of the ellipse) such that the sum of the distances from each focus to any point on the curve is identical. A circle is a special kind of ellipse in which the two foci coincide, at the circle’s centre. The further apart the foci, the more elongated, or ‘eccentric’, the ellipse. Kepler deduced that planets follow elliptical orbits, with the Sun at one focus of each ellipse (the other focus being empty). The point on an orbit closest to the Sun is called ‘perihelion’ (Greek for ‘closest to the Sun’), and the point furthest away is called ‘aphelion’ (Greek for‘furthest from the Sun’). Planets’ orbits are not strongly eccentric,and if you see them drawn in plan view they look very much like circles. For example, when Mars is at aphelion its distance from the Sun is less than 21% greater than when it is at perihelion, and for the Earth the difference is only 4%.

Kepler is justly famous for his three laws of planetary motion.Kepler’s First Law is simply the statement that each planet moves in an elliptical orbit, with the Sun at one focus. The Second Law describes how the speed of a planet varies around its orbit: a planet moves faster the closer it is to the Sun (for reasons subsequently explained by Newton’s theory of gravity) such that an imaginary line linking the planet to the Sun sweeps out an equal area in equal time. Kepler’s Third Law relates a planet’s orbital period (how long it takes to complete a circuit round the Sun) to its average distance from the Sun: the cube of the orbital period is proportional to the square of the average distance. The average distance from planet to Sun turns out to be equal to half the length of the orbital ellipse’s long axis (its ‘semi-major axis’) or, if you prefer, half the straightline distance between perihelion and aphelion.

Kepler’s laws of planetary motion enabled precise calculation of the sizes of the orbits of other planets, with an accuracy limited almost entirely by the uncertainty in how well the size of the Earth’s orbit could be measured. Even as long ago as 1672,simultaneous observations of Mars from widespread locations enabled the Earth–Sun distance to be measured as about140 million kilometres, remarkably close to the correct value of149,597,871 kilometres. Observations of the transit of Venus across the Sun’s disc in 1761 and 1769 (the latter requiring Captain Cook to station himself in Tahiti) produced a revised estimate of153 ± 1 million kilometres. Despite these and other scientifi c advances, which continued to strengthen a fully self-consistent and elegant model of the Solar System’s scale and nature, a papal ban on printing ‘heliocentric’ books in Rome remained unrevoked until 1822.

You would be excused for thinking that once the distance to a planet has been established, working out its size would be trivial.However, the smallness of a planetary disc through even a large telescope, coupled with the shimmering of the Earth’s atmosphere,leads to signifi cant uncertainty in measuring the angular size of the planet (in other words, how big it appears). For example, when he discovered Uranus in 1781, William Herschel’s measurement of its disc was 8% too large. Rather than trying to measure how big a planet looks, the most precise telescopic way to determine its size is to time how long it takes to pass in front of a star. Such‘occultations’ are rare events, but by the close of the 19th century the sizes of the planets had been determined with considerable accuracy ( Table 1 ).

Herschel discovered Uranus by accident, but Neptune was located in 1846 as a result of a deliberate search, guided by slight perturbations in the orbit of Uranus (distorting it from a perfect ellipse) that could be best explained by the gravitational infl uence of an unseen outer planet. When it had been documented for long enough, the orbit of Neptune in turn seemed to show perturbations pointing to a further undiscovered planet. This triggered a search that found Pluto in 1930. At fi rst, astronomers assumed that this newly hailed ninth planet must be similar in size and mass to Uranus and Neptune. However, by 1955 it had been shown that Pluto could be no larger than the Earth; in 1971the estimate was revised downwards to the size of Mars; and in1978 its surface was found to be dominated by highly refl ective frozen methane which meant that its physical size had to be even smaller to remain consistent with its total brightness. We now know that Pluto’s diameter is only 2,390 kilometres, so its size is smaller (and, in fact, its mass is much smaller) even than Mercury.The apparent perturbations in Neptune’s orbit that, rather fortunately, inspired the search for Pluto are now attributed to observational inaccuracies.

Table 1　The sizes of the planets (equatorial diameters)

* C. Flammarion,Popular Astronomy (Chatto and Windus,Piccadilly)

Pluto lost its status as an offi cially recognized planet in 2006.That was a contentious move, though in my opinion the right one.Before describing how this came about, I will review the nature of the Solar System as it is now understood.

A review of the Solar System

The Sun

In the centre of the Solar System is the Sun, which is a fairly ordinary star, powered by the conversion of hydrogen into helium by nuclear fusion in its core. The Sun’s diameter is 109 times and its mass is nearly 333,000 times greater than the Earth’s. It contains about 740 times more mass than everything else in the Solar System put together. Consequently, the Sun’s gravity is so dominant that objects in the Solar System orbit the Sun in almost the perfect ellipses recognized by Kepler. Perturbations to a planet’s orbit caused by other planets are tiny, though they can be measured.

The planets

Table 2 summarizes some basic properties of the planets, quoted relative to the Earth to avoid very large numbers. Distance from the Sun is quoted in ‘Astronomical Units’, abbreviated as AU,defi ned as the average Earth–Sun distance. This is fairly simple to remember as (near enough) 150 million kilometres. A planet’s orbital period is how long it takes to complete one circuit round the Sun, which is of course its own ‘year’. The orbital periods and distances from the Sun in this table are related to each other by Kepler’s Third Law. Conveniently, this means that the square of any planet’s orbital period (in Earth-years) is equal to the cube of its average distance from the Sun (in AU). The Earth’s mass is very nearly 6 million billion billion kilograms (or 6 thousand billion billion tonnes), hence the convenience of comparing other planets to the Earth rather than quoting standard scientifi c units such as kilograms, seconds, and metres.

Rotation period is how long it takes a planet to spin once on its axis. For a rapidly spinning planet, this is almost the same as the time from one sunrise to the next (the planet’s own ‘day length’),but the relationship is not exact because a planet’s orbital motion continuously changes the direction between planet and Sun. The Earth’s rotation period is 23 hours and 56 minutes, but it takes exactly 24 hours to rotate far enough to bring the Sun back to the same point in the sky. From a planet’s perspective, the Sun migrates completely round the sky during the course of a single orbit, in addition to the changing direction towards the Sun from any point of the planet’s surface caused by the planet’s rotation.A planet whose rotation had become tidally locked so that it rotated exactly once per orbit (synchronous rotation) would keep one face permanently towards the Sun. Mercury does not quite do this, but rotates exactly three times during the course of two orbits,as a result of which it turns relative to the Sun once per two orbits,so its day is twice as long as its year.

Table 2　Some properties of the planets compared. Distance from the Sun refers to average distance. Years and days are Earth-years and Earth-days. See Table 1 for sizes

There is a change in character between the four inner planets and the four outer ones. The inner planets (Mercury, Venus, Earth, and Mars) are relatively small and low in mass compared to the outer four (Jupiter, Saturn, Uranus, and Neptune). There is also a contrast in their densities, the inner planets being denser than the outer ones.The inner planets are called the ‘terrestrial planets’, signifying that they are all ‘Earth-like’. The outer four are the ‘giant planets’. Some call them ‘gas giants’ to refl ect the fact that they have so much hydrogen and helium. Others reserve that particular term for just Jupiter and Saturn, which are more gassy than the other two, though even those each contain more than one Earth-mass of gas.

Figure 1 is a map of the Solar System, showing orbits to scale,except that the orbits of Venus and Mercury are too small to include. Part of Pluto’s orbit is shown, to help with discussion later. Something that I have not yet mentioned, but without which such a map could not be drawn, is that planetary orbits all lie approximately in the same plane. Relative to the Earth’s orbit,which makes a convenient reference plane known as the ‘ecliptic’,Pluto’s orbit is inclined at 17.1°, Mercury’s at 7.0°, Venus’s at 3.4°,and all the others at less than 3°.

1.Map of the Solar System, showing planetary orbits at the correctrelative sizes. Orbits are only slightly eccentric, so look virtuallyindistinguishable from circles. The unlabelled circle inside Mars’sorbit is the Earth’s orbit, not the Sun! The orbits of Venus and Mercuryare too small to include. Pluto is not a planet, but its orbit is shownbecause it is representative of a large number of small bodies beyondNeptune’s orbit

When Pluto is near perihelion, it is inside the orbit of Neptune,but there is no prospect of them colliding. Their differing orbital inclinations prevent their paths from intersecting, and moreover Neptune is always on the opposite side of the Sun whenever Pluto passes inside Neptune’s orbit. This is possible because for every three orbits completed by Neptune, Pluto completes exactly two.Such a relationship is referred to as 3:2 orbital resonance.

As well as their orbits being nearly coplanar, every planet goes the same way round the Sun: they travel anticlockwise as seen from an imaginary vantage point far above the Earth’s north pole.Anticlockwise motion is also manifested in the direction in which each planet except Venus and Uranus rotates on its axis. Because anticlockwise motion is so common, it is called ‘prograde’.Clockwise orbital motion or rotation is regarded as backwards and is referred to as ‘retrograde’.

With the exception of Uranus, the axis about which each planet rotates is less than 30° away from being at right angles to its orbital plane. Mercury is nearly ‘perfect’, with a tilt of only 0.1°,whereas Earth’s axis is tilted at 23.5°. The direction in which a planet’s axis points and the amount of tilt both vary when measured over tens of thousands of years, but they are effectively constant on the timescale of a single orbit. Axial tilt is why planets have seasons; on Earth, summer occurs in the northern hemisphere during that part of the orbit when the north end of the Earth’s axis is tilted towards the Sun, and northern winter is six months later when the Earth is on the other side of the Sun, so that the north end of the axis is tilted away from the Sun. Of the two planets that don’t conform, Venus’s axis is tilted at only 2.7°but it rotates very slowly in the retrograde direction (giving it a day length of 116.7 Earth-days), whereas Uranus’s axis is tilted by82.1° with rapid retrograde rotation. Uranus probably suffered a catastrophe that knocked it over, having started with prograde rotation that became tipped over by 97.9° (97.9° being 180° minus82.1°). This would result in the present situation without calling on a separate event to reverse its direction of spin.

Satellites of planets

All planets except Mercury and Venus have satellites, or ‘moons’.These are smaller bodies close enough to orbit the planet rather than the Sun. Strictly speaking, a planet and its satellite each orbit their common centre of mass (or ‘barycentre’). However, planets are so much more massive than their satellites that their barycentre is inside the planet, and it is usually perfectly adequate to regard satellites as going round their planet. Most satellites’orbits lie close to their planet’s equatorial plane and almost all the large ones have prograde orbits, which is defi ned as orbiting in the same direction as the planet’s spin.

The Earth’s satellite is, of course, the Moon (with a capital M).This is exceptional in being relatively large in comparison to its planet, having a diameter 27% and a mass 1.2% of the Earth’s. By coincidence, the Moon’s size and distance from Earth are such that it appears almost the same as the Sun, which is much larger but correspondingly further away. When the Moon passes exactly between the Earth and the Sun, it hides the Sun’s disk, causing a solar eclipse. If the Moon’s orbit round the Earth were exactly coplanar with the Earth’s orbit, there would be an eclipse every lunar orbit (that is, every month). However, the Moon’s orbit is inclined at 5.2° to the ecliptic, so eclipses are rare, occurring only when the Moon happens to pass between the Earth and the Sun at one of the two points where its orbit crosses the ecliptic.Unravelling the cyclic nature of these events and predicting when eclipses would occur (though without fully understanding the reasons) was one of the great achievements of Babylonian astronomers about 2,600 years ago. Lunar eclipses, when the Moon passes into the Earth’s shadow, are controlled by the same cycle, but are more common because the Earth’s shadow is considerably bigger than the Moon.

Mars has two tiny satellites. Jupiter has four exceeding 3,000kilometres in diameter (which are those that Galileo discovered),plus at the last count 59 others less than 200 kilometres (most less than 4 kilometres) across. Saturn has a similar total number of known satellites, though only one of them rivals Jupiter’s largest. Uranus has fi ve satellites between 400 and 1,600kilometres across and 22 known smaller ones, and Neptune has one large satellite and a dozen known small ones. Most of the small (few kilometres across) outer satellites of Jupiter and Saturn were discovered using telescopes (rather than by visiting spacecraft), and more tiny satellites of the giant planets surely remain to be found, especially at Uranus and Neptune, where detection by telescope is especially hard for two reasons: they are further from the Sun and so are less brightly lit, and they are further from the Earth and so would look fainter even if they were equally well lit.

The larger satellites are geologically very interesting and I say more about them later, but all satellites are useful for the planetary scientist because they enable their planet to be weighed.The orbital period of a satellite depends only on its average distance from the planet’s centre and their combined mass (which can be calculated using Newton’s gravitational elaboration of Kepler’s Third Law). Because satellites are so much smaller, the mass of the planet dominates almost entirely, in the same way that planets’ orbits round the Sun depend on distance and solar mass.

Asteroids, trans-Neptunian objects, and comets

This book is about planets, rather than the whole Solar System,but it is worth noting that objects of other types vastly outnumber the planets and their satellites combined, although admittedly these are small and their total mass is relatively insignifi cant.Although planetary scientists have come to realize that the boundaries are somewhat blurred, these ‘junk’ objects can be divided into three broad classes: asteroids, trans-Neptunian objects, and comets.

Asteroids range downwards in size from 950 kilometres across(the diameter of Ceres, the largest example), with no lower limit.Asteroids only a few tens of metres across have been detected as they pass close by the Earth, and the remains of smaller ones that fall to the ground can be found as meteorites. Formerly assumed to be fragments of a destroyed planet, we now think that asteroids never belonged to a planet-sized object. The total mass of all the asteroids is probably less than a thousandth of the Earth’s mass.Some have clearly experienced mutual collisions, as attested by their irregular shapes.

Without exception, asteroid orbital motion is prograde. Most have orbital inclinations of less than 20°, but eccentricity is typically greater than for planets. The orbits of most asteroids lie between those of Mars and Jupiter (the so-called ‘asteroid belt’), but some come much closer to the Sun, passing inwards of the Earth and even (in a handful of examples) inwards of Mercury. A few asteroids are known orbiting beyond Saturn. Like the meteorites derived from them, most asteroids are stony or carbonaceous in composition, but some are made of iron and nickel. So far as we can tell, asteroid composition tends to be less stony and more carbonaceous and eventually icier with distance from the Sun.

Beyond the orbit of Neptune, between about 30 and 55 AU from the Sun, small icy bodies become common, and there are several that exceed the largest asteroids in size. This region is usually called the ‘Kuiper belt’, named after the Dutch-American Gerard Kuiper who predicted it in 1951 as a zone where icy lumps should be left over from the birth of the Solar System. An Irishman,Kenneth Edgeworth, said much the same in a more obscure journal in 1943, so some prefer to call this the ‘Edgeworth-Kuiper belt’. The fi rst Kuiper belt object to be discovered and recognized as such was found in 1992, but now many hundreds of them have been catalogued, and it has become clear that Pluto should be numbered among them. Similar objects with perihelion not far beyond Neptune’s orbit but reaching about 100 AU at aphelion are counted as ‘Scattered Disk’ objects. Together with the Kuiper belt,these make up a family called ‘trans-Neptunian objects’, or TNOs,all in prograde orbits. The total mass of TNOs is probably around200 times that of the asteroid belt (one-fi fth of an Earth-mass),and in total there may be nearly 100,000 bodies more than 100kilometres in size. One ‘Scattered Disk’ object discovered in 2005and subsequently named Eris seems to be marginally bigger than Pluto. We can be more confi dent about their masses because they both have satellites with well-documented orbits, showing that the mass of Eris is 28% greater than that of Pluto.

Comets have been known since antiquity because a comet can briefl y look very spectacular, thanks to the development of a tail of gas and dust that can stretch across the sky when the comet passes close to the Sun. However, the solid part of a comet is just a chunk of dusty ice (famously described as a ‘dirty snowball’), only a few kilometres across in most cases. A comet spends most of the time far from the Sun, and develops a tail only when it passes close enough for the Sun to warm it. This happens rarely because comets have extremely eccentric orbits with perihelion usually inside the Earth’s orbit but aphelion near or well beyond Jupiter’s orbit. Some come from so far beyond that their orbits look like parabolas (infi nitely long ellipses), and make only one passage close to the Sun throughout recorded history. Those are ‘long-period’ comets,and appear to have been dislodged from an ill-defi ned shell surrounding the Sun at about 50,000 AU known as the Oort Cloud. In contrast, ‘short-period’ comets probably originated as Scattered Disk objects that were perturbed into an eccentric orbit with a small perihelion distance by a close encounter with a fellow object. Those with orbital periods of hundreds of years still have their aphelion in the Scattered Disk, but aphelion can be nudged closer to the Sun as a result of a close encounter with a giant planet. For example, Halley’s comet has aphelion near Neptune’s orbit, and an orbital period of 75 years, whereas Encke’s comet has aphelion near Jupiter’s orbit and a period of only 3.3 years. Comets lose mass by degassing every time the Sun’s heat warms them, so after fewer than a thousand perihelion passages a comet is probably reduced to an inert aggregate of ice-free rock and dust, hard to distinguish from an asteroid.As you might expect, given their source, the orbits of ‘short-period’comets are prograde and tend to be close to the ecliptic. However,no such restriction applies to long-period comets whose orbits can be highly inclined or even retrograde.

What is a planet? How Pluto came to be thrown out of the club

The fi rst TNO to be discovered was Pluto, in 1930. Even after Pluto’s small size became apparent (and subsequently, thanks to the discovery of its largest satellite in 1978, its small mass), people tended to think of Pluto as the ninth planet. However, when the known population of the Kuiper belt blossomed into hundreds of objects, several of which rivalled Pluto in size, it became increasingly anomalous to classify Pluto as a planet and yet other Kuiper belt objects as something different. When Eris was confi rmed to be more massive and probably larger than Pluto,then logically either all the large TNOs had to be called planets, or none of them. However, many people argued to retain Pluto as a planet on sentimental or historical grounds.

Decision-making was hampered by the fact that the term ‘planet’had never been fully defi ned. Eventually, at a meeting of the International Astronomical Union held in Prague in 2006, during which passions ran high, delegates voted to accept some defi nitions that have largely settled the issue. Two criteria for planethood were non-controversial: the IAU ruled that, fi rstly, a planet must have suffi cient mass for its self-gravity to overcome ‘rigid body forces’ so that it assumes a hydrostatic equilibrium (nearly round) shape, and secondly that it must be in orbit about the Sun. This second criterion rules out large satellites such as our own Moon.

A third criterion was the crucial one. It states that to be counted as a planet, a body must have ‘cleared the neighbourhood around its orbit’ of everything except much smaller objects. This is the test that Pluto fails. Pluto has not cleared its neighbourhood, because it shares it with many bodies of similar size and indeed also with the vastly more massive Neptune. On the other hand, Neptune does pass the test, because it is many thousands of times more massive than anything else in the same orbital region (such as Pluto).

Having taken the bold but entirely logical step to expel Pluto from the planetary club, the IAU seems immediately to have regretted it, and invented not one but two new classes for it to belong to. At the 2006 Prague meeting, the newly coined term‘dwarf planet’ was defi ned as ‘a celestial body that is in orbit round the Sun, has suffi cient mass for its own gravity to pull it into a nearly round shape, has not cleared the neighbourhood of its orbit, and is not a satellite’. Determining whether or not shape is ‘nearly round’ is diffi cult to do remotely, and controversial to defi ne, but in adopting this defi nition the IAU gave Pluto, Eris, and Ceres (the largest asteroid) the consolation prize of being called ‘dwarf planets’. At the time, it was acknowledged that other large TNOs would be ranked as dwarf planets when they had been adequately measured, and sure enough in 2008 a Kuiper belt object named Makemake(pronounced as four syllables), estimated to be about two-thirds the size of Pluto, was judged to pass the shape test and was admitted as a fourth dwarf planet, closely followed by a fi fth called Haumea.

The IAU seemingly came to regret having lumped Pluto-like objects in the same category as Ceres, and in 2008 invented a new term, ‘Plutoid’, to denote trans-Neptunian dwarf planets. Ceres is thus the only dwarf planet that is not a Plutoid, and there is surely no undiscovered asteroid large enough ever to join it in that category. However, there are probably numerous undiscovered or inadequately documented large TNOs that will join Pluto, Eris,Makemake, and Haumea as both Plutoids and dwarf planets.Incidentally, Eris is named (appropriately, considering the controversy that it caused) after a classical Greek goddess of discord, whereas Makemake and Haumea are named after Pacifi c island fertility deities.

How it all happened

Growing planets

Until recently, it would have been possible to argue that planets are rare in the cosmos, but it now seems clear that planets are a usual by-product of star formation. The existence of our Solar System is thus a consequence of the origin of the Sun itself.

A star is believed to form when a vast interstellar cloud made mostly of hydrogen, but mixed with a few other gases and tiny solid particles referred to as dust, contracts under the infl uence of its own gravity. As the cloud contracts, most of the matter becomes concentrated into the centre, in a body that becomes increasingly hot because of the gravitational energy converted to heat by the process of infall. Eventually, the central pressure and temperature rise so high that hydrogen nuclei begin to fuse together to make helium, at which stage the central body can be called a star. The planets grow from some of the material left behind during the fi nal stages. Conservation of angular momentum causes any slight initial rotation of the cloud to speed up during contraction, and matter not incorporated into the star becomes concentrated into a disk in the star’s equatorial plane,rotating in the same direction as the star.

This rotating disk is where planets form. The one that gave birth to our Solar System is referred to as the solar nebula, ‘nebula’being Latin for ‘cloud’ and used by astronomers to denote any large mass of gas and/or dust in space. There are strong reasons for believing that the solar nebula’s composition was about 71%hydrogen, 27% helium, 1% oxygen, 0.3% carbon, and 0.1% each of nitrogen, neon, magnesium, silicon, and iron. Almost all the original dust in the solar nebula was probably vaporized by heat from the young Sun, but soon conditions in the nebula became cool enough for new dust grains to condense, not as individual elements but as compounds produced by chemical combination.Helium does not combine into chemical compounds, so the most abundant compounds that could condense involve either hydrogen or oxygen.

Thanks to the available elements and the local temperature and pressure, oxygen was able to bond with silicon and various metals to form a range of compounds called silicates in the inner part of the nebula. These are common minerals on Earth that crystallize when molten rock cools, but in the solar nebula they grew directly from gas. Hydrogen was incorporated into solid particles only where the temperature was low enough for hydrogen-bearing compounds to form, and for most purposes this seems to have been beyond about 5 AU from the Sun. At and beyond this so-called ‘ice line’, water (made of hydrogen plus oxygen) could condense as specks of ice. Further from the Sun, more volatile compounds formed where hydrogen bonded with carbon to make methane and with nitrogen to make ammonia, and carbon with oxygen to form either carbon monoxide or carbon dioxide. At about 30 AU, it was cold enough for nitrogen to condense as solid particles of pure nitrogen. By one of the quirks of planetary science vocabulary, any solid made of water, methane, ammonia,carbon monoxide, carbon dioxide, or nitrogen (or indeed any mixture of these) is referred to as ‘ice’, recognizing similarities in origin and properties. This means that, to avoid ambiguity,planetary scientists have to specify ‘water-ice’ when referring specifi cally to frozen water – a complication that rarely arises on Earth, where temperatures are too high for compounds more volatile than water to freeze naturally.

Condensation happened in such a way that the fi rst dust grains –microscopic specks made of silicates close to the Sun and ices(plus some leftover silicates) further from the Sun – did not grow as dense, rigid specks. Instead, they had intricate ‘fl uffy’ shapes,and when these bumped into each other, they tended to stick together rather than bouncing apart. Within as little as ten thousand years after the onset of condensation, the particles could have grown into globules about a centimetre across through the combined effects of continuing condensation and accretion (sticking together) when they collided. After perhaps100,000 years, the Solar System would have consisted of hordes of bodies about 10 kilometres across, dubbed ‘planetesimals’.These were all swirling round the Sun in the same, prograde,direction and enclosed in a diffuse haze made of the remaining gas and dust.

We know how long ago this happened, because some of the earliest grains survive unaltered inside meteorites. We can measure radioactive decay products within them to work out their age, which is a particularly memorable number: 4.567 billion years. The most ‘primitive’ meteorites, which are fragments of small planetesimals that never suffered heating or alteration, are called ‘carbonaceous chondrites’ and are our most direct evidence of conditions in the early Solar System.

Hitherto, collisions had been essentially a matter of chance, but once planetesimals reached about 10 kilometres in size, the greater gravitational pull of the largest ones was able to make itself felt. These suffered more frequent collisions, so their rate of growth outpaced that of the others. Within a few more tens of thousands of years, the largest planetesimals had grown to a thousand or so kilometres across, gobbling up most of the smaller ones in the process.

These large planetesimals are dignifi ed by a new name: ‘planetary embryos’. Maybe a few hundred were formed in the inner Solar System. They would have been massive enough for their own gravity to pull them into spherical shapes. They may have been hot enough internally for melting to occur, allowing iron to sink inwards to form a distinct core, but that is largely immaterial because of what happened next.

These planetary embryos were what the terrestrial planets formed from. Now that the majority of the small stuff was gone,signifi cant growth could happen only when two embryos smashed together. Such a collision is referred to as a ‘giant impact’, and liberates enough heat to largely melt the merged body formed by the collision. Imagine a sphere of molten rock, glowing red hot except for a few rafts of chilled clinker on its surface, with deep inside a ‘rain’ of iron droplets settling inwards through the silicate magma to accrete onto the central core, and you should have a picture in your mind that conveys the state of a planetary embryo in the aftermath of a giant impact.

This assumes that the impact doesn’t smash both bodies to smithereens, but inevitably a certain amount of debris will be thrown out to space as ejecta from the collision. It probably took about 50 million years to build up an Earth-sized planet by serial giant impacts between planetary embryos. Because of the randomness of the collisions and the complex ‘family tree’ of giant impact collisions between bodies that themselves had been formed by giant impacts, it is meaningless to regard any single embryo early in the process as ‘the proto-Earth’ or ‘the proto-Venus’.

Beyond the orbit of Mars, the gravitational effect of the young Jupiter was strong enough to stir the rocky planetesimals into more eccentric orbits, so that mutual collisions were often too violent to allow growth by accretion. Instead, fragmentation was a common outcome, so large planetary embryos that might have eventually collided to produce a fi fth terrestrial planet were unable to grow here. Today, in that region, we fi nd most of the asteroids, representing only a tiny fraction of the mass that once existed there. Jupiter scattered the majority into markedly eccentric orbits, so that eventually most collided with Jupiter or another giant planet, or were ejected from the Solar System entirely.

The bodies from which the giant planets formed had a high proportion of ice as well as rock within them. There, beyond the‘ice line’, the growing planets had much more material to draw upon. The role of embryo–embryo collisions is uncertain, and so is the mechanism by which they acquired so much gas. One theory is that after they had exceeded 10 or 15 Earth-masses, their gravitational pull was suffi cient to scavenge huge quantities of whatever gas survived in the remaining nebula, and so their rocky and icy kernels became encased within deep gassy envelopes.Another school of thought holds that gravitational instabilities in the nebula caused each giant planet to grow inside a particularly dense knot, where the gas was naturally confi ned about the growing planet.

Opinion is also divided over the relative rates of planetary growth in the inner and outer parts of the Solar System, and it is unclear whether Jupiter formed before or after the Earth and Venus.However, if they grew by embryo–embryo collisions, Saturn,Uranus, and Neptune must have grown more slowly than Jupiter because collisions should be less frequent with increasing distance from the Sun.

Scavenging of gas from the nebula was terminated when the Sun entered its ‘T Tauri’ phase, named after the star T Tauri which is undergoing this process today. For perhaps 10 million years, a strong outfl ow of gas from the star, called the ‘T Tauri wind’, blows away all the remaining gas and dust. A likely reason for Uranus and Neptune having proportionally less gas than the other giant planets is that they took longer to grow, leaving less time to collect gas before the T Tauri wind put an end to the process.

Migrating planets

A further matter of debate concerns ways in which orbits can change over time and the extent to which this happened,particularly among the giant planets. Until the solar nebula was dispersed, gravitational interactions between nebular material and large orbiting bodies would tend gradually to decrease the radius of their orbits, causing embryos and young planets to migrate inwards. After nebular dispersal, gravitational interactions between planets and smaller bodies could have played an even more dramatic role. Some suggest a period of half a billion years or so when the outermost giant planet was defl ecting the orbits of outlying icy planetesimals inwards, where they would be likely eventually to be nudged further inwards by interaction with the next giant planet, and so on until they passed close enough to Jupiter for Jupiter to fl ing them outwards. These out-fl ung icy planetesimals could be the origin of today’s Oort Cloud. Jupiter must have moved fractionally closer to the Sun each time it fl ung a body outwards, but conversely the other giant planets would have been nudged outwards each time one of them swung a lump of ice inwards.This story has Jupiter migrating inwards, while Saturn, Uranus,and Neptune migrated outwards. It is even possible that Uranus and Neptune swapped places (providing an opportunity for Uranus’s axis to become tipped over into its present state).Today’s TNOs are those that survived beyond the zone swept clear during Neptune’s outward march.

Please do not form the impression that the orbit of a planet is capable of changing either rapidly or dramatically. Claims that Venus and/or Mars passed close to the Earth during biblical times,triggering various myths and natural disasters, are completely untenable. The outer planet migrations I have described happened extremely slowly, and as a result of cumulative interactions with nebular gas and with vast numbers of small bodies that are no longer available.

Nevertheless, the planets and their mutual gravitational pulls are continuously changing confi guration. Chaos theory says it is therefore impossible to predict planetary positions more than a few million years ahead. However, it can be shown that the Solar System is suffi ciently stable that no planet is likely to collide or be ejected in the next few billion years. We are probably safe for at least 5 billion years, which is when astronomers expect the Sun to swell up into a red giant, whereupon the wanderings of Mars will be the least of the problems faced by any far future Earthlings.

Why all the satellites?

By now, you should not be surprised that there is no straightforward answer to the question of whether satellites somehow grew alongside their planets or were acquired later.The large prograde satellites of the giant planets are the easiest to explain. They are thought to have formed within a cloud of gas and dust surrounding each giant planet as it grew, rather like a miniature version of the solar nebula. Tiny prograde satellites only a few kilometres in size orbiting close to giant planets are probably fragments of larger satellites that came too close and broke apart. The outer satellites of giant planets are mostly in retrograde orbits, and these are probably captured bodies that began as asteroids, TNOs, or comet nuclei.

Theoretically, it is almost impossible for a planet to capture a passing visitor into orbit about itself. An incoming smaller body will be swung past a planet by the pull of its gravity, but it can’t easily be slowed down enough to be captured into orbit. However,if the visitor is a double object, one of the pair can be captured by transfer of momentum to the other member, which will scoot away even faster after the encounter. A currently favoured explanation for Neptune’s large retrograde satellite, Triton, is that Triton was formerly half of a double TNO that strayed close to Neptune. This seems plausible, because several TNOs are known to be twin bodies. Mind you, it leaves unresolved the issue of why so many TNOs (and indeed asteroids too) have satellites in the fi rst place.

The Earth’s Moon has a different explanation, and seems to have condensed from debris thrown into orbit about the Earth by the fi nal embryo–embryo collision of the series by which the Earth grew. The two tiny satellites of Mars (Phobos and Deimos) are asteroids, whose capture into close circular orbits is not understood.

Collisions and the cratering timescale

Although collisions between substantial objects are now extremely rare, there is still a large number of small objects that could eventually collide with a planet. Until about 3.9 billion years ago(an epoch called the ‘late heavy bombardment’), the rate at which asteroids and comets were hitting planets was much higher than today. Impact craters of that age are well preserved on the Moon( Figure 2 ), though cratering has continued at a slower rate ever since. An impact crater forms on a solid body when something hits it at a few tens of kilometres per second, and is excavated by shockwaves that radiate from the point of impact. Craters are circular, except for rare examples when the impacting body arrives at a grazing angle.

There is a well-understood hierarchy of crater morphologies depending on diameter, and which can be reproduced experimentally and in computer models. On the Moon, craters from microscopic size up to 15 kilometres across have simple bowl shapes.Then up to about 140 kilometres diameter, craters do not become deeper but have fl at fl oors, and usually a central peak formed by rebound immediately after excavation. There is a nice example near the top of Figure 2 . Larger craters may have a group of central peaks,and then craters bigger than 350 kilometres take the form of two or more concentric rings. The transitions from one type to another occur at slightly smaller diameters on bodies with stronger gravity.

Earth’s cratering record is poorly preserved, because it is an active planet where processes that erase or bury craters almost keep pace with the rate at which craters are formed. Fortunately, the vast tracts of ancient terrain surviving on the Moon allow us to count the density of impact craters on surfaces whose ages are known,thanks to datable samples returned to Earth by the Apollo programme of manned lunar landings, supplemented by a few Soviet unmanned sample-return missions. By this means, we know the date of the late heavy bombardment and also the average rate at which cratering has affected the Moon ever since.The Earth must have been exposed to the same fl ux of impactors as its satellite, and there are good reasons for believing that this is also a reasonable approximation for Mercury, Venus, and Mars.Counting craters is thus the best way we have to estimate ages on planetary surfaces. Even if the absolute age is in doubt, it is usually safe to assume that a surface with a lower crater density is younger than one with a higher crater density.

2.A 470-kilometre-wide view of a heavily cratered region of the Moon.Most of these craters date from about 3.9 billion years ago, obliteratingany older craters. Each crater was formed by the impact of a bodyabout 20 to 30 times smaller than the crater itself. Parts of the Earthwould once have looked similar

At present, the Earth is hit annually by about 10,000 meteorites greater than 1 kilogram, but most of those are too small to survive passage through the atmosphere, where they are heated and worn away by friction. The yearly supply of 1,000-kilogram meteorites is only about 10, and the average interval between impacts by150-metre-diameter objects (which would produce a crater some 2kilometres across) is about 5,000 years. Impactors about1 kilometre in diameter arrive at random about once every200,000 years, boring through the atmosphere as if it were not there, hitting the ground with undiminished speed, and forming a crater perhaps 20 kilometres across. Larger, and more devastating,impacts are even less frequent.

Collisions affect each body in the Solar System, but craters survive only where there is a solid surface and insuffi cient other activity to erase the record. Observers were fortunate to discover a string of fragments of a tidally disrupted comet shortly before they collided with Jupiter in July 1994. Several of the collisions were witnessed,and each left a brown scar in the giant planet’s atmosphere that lingered for several weeks, as did a scar found in July 2009 made by an unobserved single impact.

Planets as abodes of life

If Earth were not at a comfortable distance from the Sun, you would not be reading this book, because life may not have become established and we could not have evolved here even if it had.Scientists talk of a ‘habitable zone’ around every star, at a distance where the surface temperature on a planet would be neither too hot nor too cold for life. By analogy with Goldilocks’ preference for baby bear’s porridge (whose temperature was ‘just right’), the habitable zone is sometimes called the ‘Goldilocks zone’. In this context, ‘habitable’ means somewhere that could sustain life of any kind, even just simple microbes. It does not imply that the environment would be inhabitable by humans.

Because our kind of life requires water, the habitable zone is usually equated with the distance from a star at which the surface temperature of a planet would enable water to exist in liquid state.The density and composition of a planet’s atmosphere infl uences the surface temperature, but the main control is heat received from the star. The habitable zone around the Sun is estimated to extend from about 0.95 to about 1.5 AU. These estimates put Venus (0.72 AU) well inward of the inner edge of the habitable zone, and Mars (1.52 AU) at its outer fringe. The Sun’s output has probably increased slightly since the planets were formed,nudging the habitable zone outwards over time, so Mars would seem to be a poor candidate for life, but it is not a hopeless case.

The idea of a habitable zone defi ned by planetary surface temperature has been criticized as too narrow. There are circumstances where heat generated within a planet may provide an environmental niche suitable for life, although the surface may seem inhospitable. Even on Earth, we know of ‘extremophile’organisms living below 0 °C or above 100 °C. Thus, even if all life is, like that on Earth, based on chains of carbon molecules and dependent on water as a solvent, there are several places in the Solar System where it could exist (though only one, the Earth,where it is known to exist) and at least many millions of habitable places elsewhere in the galaxy. I shall return to that theme near the end of the book.

Space exploration

Telescopes are very useful, for example to measure the temperature and composition of a planet’s surface and atmosphere. The polar ice caps on Mars were correctly identifi ed by William Herschel as long ago as 1781. Jupiter is suffi ciently large and close that storms among its clouds can be monitored even with fairly modest telescopes. However, this book would be duller and more speculative were it not for half a century of space exploration when space probes from Earth have visited every planet in the Solar System. Soviet probes reached the Moon in1959, and twelve American astronauts walked on its surface between 1969 and 1972. Unmanned American (NASA) and Soviet probes fl ew past Venus and Mars in the 1960s, and achieved orbit and soft landings during the 1970s. The fi rst Jupiter and Saturn fl y-bys were in the 1970s, and the other giant planets were visited in the 1980s. Since 1990, the terrestrial planets have been explored by increasingly capable orbiters, robotic rovers have crawled across the Martian surface, and complex orbital tours of both Jupiter and Saturn have been achieved.

The most famous missions include Vikings 1 and 2 that landed on Mars in 1976; Magellan that mapped the surface of Venus by radar 1990–4; Voyagers 1 and 2 that fl ew past the giant planets between 1979 and 1989; Galileo that orbited Jupiter between1995 and 2003; and Cassini that began an orbital tour of Saturn in 2004 and sent a probe named Huygens to the surface of Titan in 2005.

Highlights in the years ahead include return to Earth of samples collected from Mars, asteroids, and comets, and a renewal of human presence on the Moon. The USA and Russia are no longer the only space powers. The European Space Agency has gone solo to Mars and Venus, to Saturn jointly with NASA, and will soon go to Mercury with Japan. The Japanese have sent probes to the Moon and to an asteroid, and China and India have each reached the Moon. Scientifi cally, there has been much cooperation (and most probes carry instruments contributed by several nations),but it is undeniable that there is also national pride at stake,together with long-term strategic and commercial interests.