Sea Room. Adam Nicolson

of shrinking material, contracting over its entire width, pulls apart from itself internally. Shrinkage cracks develop in the body of the rock identical to the network of polygons that develop on the floor of a drying lake. The columns are nothing more than a network of cracks extended into a third dimension.

      If this cooling had been conducted in laboratory conditions, where the magma sheets were of equal thickness throughout and both upper and lower surfaces of the intrusions were level, then a structure of complete regularity would have emerged. All the columns would have been straight, the same size, and parallel. But this is not a laboratory. Some of the intrusions were clearly thin (in particular the one that formed the Galtas) and cooled more quickly. This has meant that the columns are themselves much thinner there. More intriguingly and beautifully, it is clear that the opening into which the magma squeezed was uneven. The columns would have formed by growing perpendicular to the cooling surfaces and here the unevenness of those surfaces has created columns that curve, twist and bend, are waved like the hair of art deco statuettes, fixed in the elegance of a geological perm, Jean Harlow turned to stone, Madonna having glimpsed the Gorgon.

      Others, such as the upper sections of the north cliffs on Garbh Eilean, cooled so quickly that whole slaggy masses of rock became solid before columns could develop. Below them, deeply buried in the huge Garbh Eilean sill, the magma cooled very slowly indeed and here the Shiant rock-forms attain the great magnificence of the giant columns. These were the forms over which John Macculloch, the early geologist, enthused in 1819:

      The lover of picturesque beauty will here, as in many other parts of the Western islands, be gratified with a display of maritime scenery combining the regularity of Staffa with the grander features of the coast of Sky. Towards the north it exhibits one continuous perpendicular face of naked rock. This face is columnar throughout, and forms a magnificent scene for the pencil; spreading in a gentle curve for a space of 1000 yards or more, and impending in one broad mass of shadow over the dark sea that washes its base. In simplicity and grandeur it exceeds Staffa almost as much as it does in magnitude; offering to the tourist an object as worthy of his pursuit as that celebrated island, and of no very difficult access from the northern extremity of Sky.

      The Shiants languished in obscurity, while their more famous, and more accessible cousin-rocks off Mull became ever more visited. Perhaps the open waters of the Minch protect the Shiants from fame. Perhaps the nearness of Staffa to Iona creates its public success. And having witnessed, from the deck of a boat, Staffa sagging one summer’s day under the weight of its geological trippers, I can only say; ‘Thank God’.

      A Gothic fate awaits the Shiants. There has been a steady geological drizzle over the millennia which has created the huge scree slopes at the feet of the cliffs. Giant pencil stubs the size of small houses lie tumbled like the aftermath of an earthquake. Sometimes, groups of them still hang together as if thrown in a clump on the rubbish heap. The birds live in many-storeyed tenements among them. If you walk across them, the lorry-sized rocks wobble and creak beneath you. Shiant dynamism is not over. The cliffs themselves are a symptom of the slices being taken out of them and the hard edges of the islands are signs of destruction in progress. ‘The upper millstone heaven,’ Ted Hughes once wrote, ‘Grinds the heather’s face hard and small.’ It isn’t only the heather. Fergus Gibb reckons that ‘a million years or two should see the Shiants off.’

      Every spring, I look for the new scars, the beds from which the lumps of rock have broken away. They are unnerving places. Where the splits have occurred, the remaining edges are as sharp as knives. You can cut your hands on them. For some reason, the bare unlichened stone smells of iron or even blood, because blood smells of iron too. The smell is one of deep antiquity, a release into the nostrils of elements in the rock which have not been volatile since the rock was made. It feels as intimate as poking your fingers into a wound.

      I have never witnessed something which I have spent hours in a boat waiting to occur: the collapse of an entire column from a cliff. Fergus has only seen it once, and then not here but in the similar rock formations in Trotternish, the northern wing of Skye, twelve miles or so to the south. He too was in a boat on the quiet sea. Alerted by a shuffling, a distant rumbling in the silence which on still summer days hangs around these places, Gibb looked up from his notes. Across the bay, an entire thin pencil, perhaps three hundred feet high, six or eight feet across, was slipping in slow motion into the sea. The base of the column, like many of them, must have been eaten away by the sea. Incredibly, the columns of which the islands are made are scarcely more bound to each other than pencils in a box and once the base has gone, knocked out by a winter storm, there is nothing to withstand the force of gravity. That morning, the column slid down, buckled and then fell, not like a felled trunk but with the shaft snapping in two places in mid-air before the three giant sections crashed like stone hail into the stillness of the Minch. The birds clattered away from the impact, the wash ran up to Gibb’s boat and on past it and the silence pooled back in. Fergus said it was like a glacier calving.

      These rocks are killers too. In 1796 the Reverend Alexander Simson, the Minister of Lochs, described the Shiants in his statistical account of the parish:

      There is one family residing on the largest of the islands. The head of this family has been so unfortunate as to lose, at different times, his wife, a son, and a daughter, by falling down great precipices; the mother and son met with this catastrophe in following sheep, and the daughter, by going in quest of wild-fowl eggs.

      There is no further explanation of why they should have fallen. It is easy enough to slip on the dew-wet cliff-top grass, or to be blown away in a sudden gust, or for a rope to fail, but it seems likely enough to me that the collapse of part of a cliff might be to blame.

      More recently, the sheer instability of Shiant rocks killed a boy. On 28 June 1986, a party of teenagers and their teachers from Cranbrook School in Kent had just arrived for a summer expedition to the islands. It was a beautiful evening, and as they were putting up their tents, one of the boys, Simon Woollard, an experienced alpinist and gifted climber, decided to climb the small cliff just behind the house. It is no more than twenty feet high and I have often climbed it myself, pushing up past the bunches of wild thyme, the purple knapweed and the hart’s tongue ferns, without ropes, for fun. We had given names to some of the routes – ‘Grassy Chimney’, ‘The Squeeze’, ‘Crab Lunch’ – ‘a naughty little climb with pretentious reaches’, as I wrote in the visitors’ book when I was seventeen.

      That evening in 1986, Simon Woollard did it by the book. He was belayed from below, wearing a harness and a helmet. His friends were watching him from among their tents on the grassy level behind the house. The Shiants on a summer evening like this, as the sun begins to drop towards the hills of Harris, and the two Galtas stand out as a pair of black moles against the colours of the evening, and as the birds come in from their fishing for the evening wheel between the islands, is the happiest and calmest of places. It could not hurt you. It was then, at about twenty to nine in the evening, that, towards the top of the cliff, a block of dolerite over which Simon was pulling himself came away in his hands. It was about the size of an armchair. He fell with it for a moment but was then held by the rope and the rock sliced through his helmet and into his head. He died there, at the foot of the cliff where he had fallen, and later that night, after many hours’ delay and unspeakable distress for those who were there, a helicopter from RAF Lossiemouth came to take his body and some of his friends back to Stornoway. The others left thirty-six hours later, when Donald MacSween collected them from Scalpay. There is a small plaque at the place where Simon died and none of us has ever climbed there again.

      That is not quite true. I came to the islands a couple of weeks after Simon Woollard had died. The evidence was still there: shards of broken rock on the turf at the cliff foot, still sharp, a huge and horrifying stain on the boulders which the rain had not yet washed away. On the evening of the accident, Adam Tozer, the master in charge, had written in the visitors’ book that there had been a fatality. With big, slashing, diagonal lines he had crossed out the pages in which we had described the various routes. ‘DO NOT CLIMB’, he had written across the sketch of the rocks on which the boy had died.

      For the first day or two I did what he