By about 3 million years, Jupiter and Saturn had formed and were cooling down. But the protoplanetary disc was still very active. Closer to the Sun, the rocky planetesimals were continuing to gather. And much further from the Sun – twice as far out as Saturn is, and beyond – so too were the last of the icy planetesimals. Despite the abundance of ice there, it took longer for icy protoplanets to accrete to the dimensions where, like Jupiter and Saturn, they could pull in gas directly from the disc, because the orbital speeds there were slower. Eventually, though, two more dominant protoplanets of ice and rock did develop. These would become the outermost giants, Uranus and Neptune.

In time these kernels of rock and ice, each about as massive as the modern Earth, began to stockpile hydrogen and helium, just as the larger cores of the gas giants had done a couple of million years earlier. But they had arrived on the scene too late. The Sun was by now past its T-Tauri phase, and very little gas remained in the protoplanetary disc. For a few more million years Uranus and Neptune seized what little gas they could from the ever-diminishing supply, but their growth ceased after about 10 million years – the exact time remains uncertain. The end result was a pair of planets a little over one-third the diameter of Jupiter and only 5 per cent of its mass. And yet, despite their diminutive statures compared with Jupiter, Uranus and Neptune are each still heavier than 15 Earths. They were more than capable of joining in the game of cosmic billiards demonstrated earlier by Jupiter and Saturn.

The terrestrial planets were latecomers. Because ices could not condense near the Sun, the materials (rock and metal) from which these planets coalesced were a lot less abundant than those that formed the giants further out. So, while the gas planets had formed within a million years – or at most a few million years – and the ice giants took maybe ten million years, for the terrestrials the formation process was even longer.

At least the initial growth of the terrestrial planets, within a few astronomical units of the Sun, had been very fast. Once the first rocky planetesimals had appeared, they had begun gravitationally to attract smaller bits of nearby debris. As we have seen, these first planetesimals grew to dimensions of hundreds or thousands of kilometres in less than 100 000 years. After about one million years the innermost regions of the Solar Nebula were populated by several large rocky and metallic protoplanets approaching the size of Mercury. And by 10 million years these protoplanets had grouped together through gravitation so that only four dominant spheres remained. These, at last, were the primitive terrestrial planets: from the Sun outwards, Mercury, Venus, Earth and Mars. But even after all four of the giants and their satellites had emerged, the terrestrial planets had grown to only half their eventual masses. And they had a very long way to go to make up that missing half – because the supply of available fragments in the disc was now much lower. Moreover, the terrestrial protoplanets had become large enough for the addition of more planetesimals to have a smaller and smaller effect on their size as they continued to accrete.

While the four giant planets were forming, they were not doing it alone. As each of the giant protoplanets stole gas from the Solar Nebula, the material had swirled around the icy kernels to form gas discs like the Solar Nebula on a much smaller scale. Exactly as in the Solar Nebula itself, the particles in these discs had begun to lump together into larger building blocks – and new, independent worlds had started to appear in orbit around the planets. These would become the giant planets' satellite systems – their moons. Because these moons formed from discs, like the planets, they now tend to orbit their planetary hosts in a thin plane, each in the same direction as the others and in fairly circular paths. Moons with these orbital characteristics also tend to be large. They are known as regular satellites.

It is probable that the regular satellites grew to maturity very quickly, even before their planets did. Why? Simply a question of scale. The discs that surrounded the newly emerging giant planets were much smaller than the Solar Nebula, so they had correspondingly shorter orbital timescales. Their rich cargoes of icy volatiles grew to protoplanet dimensions much more quickly than the planets did. But not all of the moons formed at the same time. The Jovian disc, right on the snow line, would have been the richest. So Jupiter's regular satellites – Io, Europa, Ganymede and Callisto – no doubt formed first, alongside their planet, at T-plus 2–3 million years. These are known today as the Galilean moons, after their discoverer.

Thirty million to 50 million years. That's all the time it took to form the star we call the Sun. This may sound like a long time, but let's put it in perspective. Since the last dinosaurs walked the planet, enough time has passed for at least one and possibly two stars like the Sun to have formed, one after the other – utterly from scratch. The details of this miraculous creation are not exceptionally well understood, but astronomers at least have a good grounding in the basics. Perhaps ironically, one star's birth starts at the other end of the line – when other stars die.

Generally speaking, stars make their exit in one of two ways. A low-mass star like the Sun eventually expands its outermost layers until the star becomes a gross, bloated caricature of itself: a red giant. Gradually, the star's envelope expands outwards, all the time becoming thinner, until the dense core of the star is revealed. Such an object is known as a white dwarf. It is a tiny and, at first, white-hot object with a stellar mass – yet confined to live out the rest of its existence within the limits of a planet's radius. The rest of the star meanwhile, the cast-off atmosphere, grows larger and larger. Eventually it becomes nothing but a thin fog of gas spread over more than a light-year. This is the fate that awaits our Sun, as we shall see in detail in Part 4. By contrast, a heavier star dies much more spectacularly. It blows itself to smithereens in a star-shattering explosion called a supernova.

After about 10 billion years, the surface of the white dwarf Sun cools to around 3000–4000 Celsius. At these temperatures the object looks distinctly red (even though it is still called a white dwarf), and is tens of thousands of times dimmer than the main-sequence star it used to be. By now the cooling rate, which slows drastically with age, is incredibly slow. But though it takes an unfathomable amount of time – longer even than the current age of the Universe – the white dwarf Sun one day vanishes totally from the optical window in the electromagnetic spectrum through which we humans today admire the Universe. Too cold to emit any signs of optical radiation at all, the dead Sun ceases to shine. It becomes a black dwarf.

At long last, perhaps 100 billion years from now, maybe even longer, the light will go out in the Solar System. The battered planets still remain, their orbits being stable, huddled around a star that they can no longer ‘see’. But overall, the scale of the planetary realm is almost twice as large as it is today. At only half a solar mass, the dark star's gravity clings quite feebly to its retinue of charred worlds, and they each orbit about 1.85 times further out than they do today. Meanwhile, those planets look little, if anything, like the worlds we know today. Over the tens of billions of years, facing a steadily declining heat source, the terrestrial planets have cooled down to just a few degrees above the coldest temperature possible, absolute zero.