I had this picture on the wall of my room at school and it remains a consoling image for me. Daniell does not attempt any heroics, any wild dynamism in the picture. He portrays the islands as a place of quiet, with a glow in the light and the huge, brooding stability of the cliffs behind them.

      It never lasts. The Atlantic drives its next weather system on towards the islands and that sullen, lit beauty is taken up and twisted into a new and familiar frenzy. The Shiants’ temper is like a child’s: unbidden, unexplained rage; sudden quiet; a new paroxysm as total as the one before. Awake at night in the house, I lie listening to the weather. I see the Shiants as if from above, laid out beneath the storm. The cloud shadows beat across them. The swells cram themselves, one after another, through the natural arch, filling it, forty feet high and thirty wide, a tube of white water a hundred yards long, squeezed in there, until they burst out on the far side, released into huge, disintegrating flowerheads of surf. The cliffs and the islands are unmoved. Besieged by the Minch, they remain there, black, impassive and irreducible.

      It is tempting to see the Shiants in that way. Perhaps any island owner would like to think of his property as a hedge against erosion but it couldn’t be more wrong. I once spent a few days on the Shiants with a pair of geologists and under their steady rational analysis all idea of the island fortress was soon whittled away. It was a highly enjoyable experience. Fergus Gibb, the Reader in Igneous Petrology at Sheffield University, and his friend Mike Henderson, now Research Professor of Petrology at Manchester University, both know more about the Shiant rocks than anyone on earth. They have been studying the islands since the 1960s and in a small dinghy they guided me around the cliffs and shore, pointing out to me where the story of the Shiants was to be found.

      It was a charming, affectionate and mutually impatient double act. Fergus – Mike calls him Fergie – is the more bullish and macho of the two. He plays tennis for the Yorkshire Veterans, talks with fervour about ‘stonking great sledge-hammers’, likes to give things ‘welly’, wears dark glasses and short-sleeved tartan shirts, and looks after Mike, whose balance on the rocks is uncertain. Fergus’s big seamen’s stockings are always pulled up over his trousers to his knees. He takes charge and one falls into line. Mike – knitted wool tie, glasses, green V-neck, baseball cap well down on the brow – plays the complementary role. He and his wife Joan are keen on organic food, a certain kind of witty late nineteenth-century novel and the finer of the performing arts. Mike must be the only geologist in the world who has had a ballet dedicated to him and the first thing he asked me, as I sat him down to a supper of roast lamb in the house on the Shiants, was what I thought of Giselle.

      Neither talks of immovability or irreducibility. They are engaged with something richer and deeper: the huge, slow dynamism of this extraordinary place. Everything in the geologist’s mind is a symptom of something happening. We layman landscapists may see the thing itself, the immovable rock, the huge columns, the stand of swaying bamboo, the Elizabethan ruff, the clustered organ pipes; they see the process, the mineralisation, the conductive cooling, the developing faults. ‘Don’t think of what it is,’ Fergus said to me. ‘Think of how it came to be.’

      The Shiants, or at least most of them, are about fifty-eight and a half million years old. They are formed from a series of hot, intrusive magmas, giant plugs of molten rock rising from deep within the Earth’s mantle, which squeezed between much older fossil-bearing rocks above them. The process is about as dramatic, and as unlocal as it could be. The ‘emplacement’ of the magma was part of an event of planetary scale. About sixty million years ago, a whole zone of the Earth’s crust began to come under immense strain. Whether the weakness in the crust caused the upwelling of hot rock from deep below; or whether the upwelling of the magma, a huge bubble of heat and energy in the mantle, caused the weakness in the crust, is not certain. What is sure is that this was a stretched, tensed time. Heavy convection currents in the mantle, rising in a plume beneath this spot, put the whole of the Hebrides under pressure. A boil covering half a continent wanted to burst. The whole depth of the Earth’s crust was being stretched here, one part being pulled south-west, the other north-east. The result was one of the most cataclysmic episodes in the history of Britain. The precise geography of this is still not quite clear but it is certain that the enormous volcanic outpourings would have been visible from another planet. Signs of the rift system run from Disco and Nunavik on the west coast of Greenland, to Kangertittivaq on the east coast, out to Jan Mayen Island, deep within the Arctic Circle, besieged by drift ice, across tens of thousands of square miles of what is now the North Atlantic, through Iceland, the Faeroes, Rockall and St Kilda, down through the Shiants, all the way through Skye and the Inner Hebrides, to Staffa and the Giant’s Causeway in Antrim, on to Slieve Gullion and the Mountains of Mourne, before ending in the granite of Lundy in the Bristol Channel.

      The rift may perhaps have been the first attempt at the opening of an Atlantic Ocean. If so, the Hebrides would have been pulled apart. Floating in Freyja between the islands, I am directly above one of the world’s great might-have-beens. From within the centre of the Shiants, America and Europe would have moved slowly away, eased apart on giant conveyor belts of hot rock in the mantle. The Galtas and Garbh Eilean would now be off the coast of a New England shore made principally of the old and twisted rocks of Lewisian gneiss which form the Outer Hebrides. Eilean Mhuire would be the westernmost island of Europe, looking out not to the screes and green puffin slopes of Garbh Eilean, a mile away, as it now does, but across three thousand miles of grey Atlantic.

      It didn’t happen. The rift never opened beyond a slit. The Atlantic opened to the west of Rockall and the Hebrides remained whole. All that is left of this cataclysmic episode now are the roots and the remnants. Nearly sixty million years of erosion has done its work and almost nothing of the surface landscape which this vulcanism produced now survives. The great mountain punctuation points of the volcanic province, Slieve Gullion and Mourne themselves, the hills of Mull and Ardnamurchan, the Cuillins of Rum and Skye, represent only the hardened footings of the enormous volcanoes through which this spasm of the Earth’s intestinal juices were vented on to the surface.

      The Scotland into which the lavas poured was tropical and paradisical. America, Britain, and the Europe of which it was a part, would have been unrecognisable then. Lotus lilies, magnolias and several species of proteus, the tender plant which now flourishes in South Africa, were all thriving in the Hebrides. After each eruption there was a pause. Earth and life accumulated on the ragged, fissured surface of the lava before the next eruption destroyed and enclosed it. In Morvern, one can see the upper parts of each flow stained red where the weather had broken down the rock and turned the iron in it rusty. Each of these red layers is buried under the basement of the flow that came after it. On Mull, at Rudha na h-Uamha, John Macculloch, the fiercely opinionated geologist who was one of the first to give an account of the Shiants’ rocks in 1819, discovered the fossil of a tree still standing embedded in a river of lava, twenty feet deep, which had overwhelmed it. You can still see it there today.

      The Shiants themselves, though, can never have known any of this. They never had their Hawaiian period and parakeets never flitted along these shores. These islands – or what are now these islands – came into existence two miles or so underground, perhaps under another volcano which has entirely disappeared.

      That much was known in the nineteenth century. Since World War II, the understanding of the Shiant rocks has developed enormously, largely in the hands of Fergus Gibb and Mike Henderson, who, in five visits and with extensive laboratory analysis, have, quite extraordinarily to my mind, been able to establish a sequence of events that occurred here over a few decades about sixty million years ago.

      It is a vast poem written in heat and liquidity. Remove the sea from your mind. That has nothing to do with it. You are in the dark of the Earth’s crust, several miles down. About a hundred and twenty million years before, in the Lower Jurassic, enormously thick accumulations of mud, silt and sand had built up in layer after layer on the floor of an earlier sea. These mud stones and shales have become rock. Beneath them, a huge bubble of hot molten