Glaciers and waterwheels have blocked the path of the salmon, our oldest native vertebrate. Can it, yet again, stage a comeback, asks Charles Rangeley-Wilson

In an extract from his book, Silver Shoals: Five Fish That Made Britain, Charles Rangeley-Wilson considers what has caused the decline of Atlantic salmon, and whether our oldest native vertebrate can stage a comeback.

Learn about the history of the River Wye and how it garnered a world-wide reputation for enormous salmon. Read River Wye: the Wye and the wherefores.


I read an interview many years ago with the palaeontologist Richard Fortey, in which he perfectly summarised the sibylline link between geology, natural history and human history. “Far from being the driest of sciences,” Fortey said, “geology informs almost everything on our planet and is rich with human entanglements. The rocks beneath us are like an unconscious mind beneath the face of the earth, determining its shifts in mood and physiognomy.”

The natural and unnatural history of our wonderful but beleaguered Atlantic salmon has been greatly determined by geology and the forces of the Earth’s evolution, suggesting one hopeful thought – that somehow the salmon will survive – and one terrible thought – that even so we may yet lose them from the British Isles.

Knowing exactly why the salmon is in such decline remains a Holy Grail of fisheries science, and the driving force behind much urgent research. But the shaping influence of that deeper geological history is not only fascinating, it may offer useful insights into how we hold on to what we’ve got.

Decline of Atlantic salmon

Atlantic salmon leaping a weir on the River Tyne in Northumberland during their upstream migration.

William Smith drew the first and still the most beautiful geological map of Britain. In marbled swirls of pink and purple he mapped the bands of igneous and metamorphic rocks which stretch all across the south-west and up through Wales, north-west England and into Scotland, while in shades of blue, yellow and green he mapped the progressively younger and softer sedimentary rocks which cover southern and eastern England. There’s a discernible change from one spectrum to the next and this division, which marks the transition from Triassic to Jurassic also marks an approximate change in topography between a more generally upland and a lowland Britain.

Overlay onto Smith’s map all the rivers that still hold salmon today and the border between their presence or absence would almost exactly trace that division. In other words, the rivers that still contain salmon today are mostly upland and rise from harder, older rocks, while the blue-green counties and coastlines of Sussex, Kent, Essex, Suffolk, Norfolk and Lincolnshire, the Home Counties, most of the Midlands and the east coast as far as North Yorkshire are, apparently, a salmon desert. Only on the very edge of the green are there any post-Jurassic sedimentary streams which salmon run – the chalkstreams of Hampshire, Wiltshire and Dorset – though they are only just holding on. The Piddle and the Stour reported no fish at all in 2015. The other chalkstreams between them reported just 781.

You might suppose from this that salmon just don’t get on well in the rivers that drain those sedimentary landscapes of the south and east, that with the exception of the Thames, salmon were simply never there in the rivers of lowland England. But that makes no sense at all. Trout, the salmon’s very close first cousin, thrive in these chalk, clay and greensand rivers. The Thames – which once produced salmon in such numbers as no other river in Europe could rival – is fed by dozens of tributaries that are just the same and these are where the Thames salmon spawned: rivers like the Windrush, Evenlode, Kennet and Pang. If salmon ran these, there’s no geological or ecological reason why salmon would not once have run similar rivers all across the south and east. Even if they don’t run them now. Even if the pattern of their absence appears to be geologically determined.

Geology being ‘rich with human entanglements’ there will be other maps that could be revealingly overlaid – land-use, for example – but one of the most revealing is that of early medieval watermills. Margaret Hodgen drew one for her 1939 paper The Domesday Watermills, published in Antiquity. In 1086, she revealed, William’s auditors counted 5,264 mills in his newly conquered kingdom. And just like sedimentary rocks, these mills when marked on a map are also mutually exclusive with present-day salmon either side of that geological divide along the edge of the Triassic.


Medieval watermills. Twenty-first-century salmon. They are bound together in mutual opposition by geology because the lowland landscapes of the south and east are more malleable, homely and fertile than those to the north and west, and because their rivers are spring-fed and ideal for the harnessing of water power.

It’s not as if there were no watermills in the harsher landscapes to the west and north, but the land was not as fertile so there were far fewer, spread across many more miles of river and designed – because of geology – in such a way as to not impede a migrating salmon. Northern mills, where they existed, tended to be of the Norse type, constructed off the main rivers – to protect the buildings from being destroyed in the frequent and violent floods that are characteristic of spatey, rain-fed rivers – with a mill-wheel that sits in the flow and requires no gearing.

In the south and east mills were built right across the stream because they could be built right across the stream. To get the best out of the constant but gentle flows of a lowland stream the Domesday mills were of the ‘Roman’ type, where gearing is used to flip the vertical rotation of the waterwheel into a horizontal rotation of the grinding stones. A Roman mill works by impounding the river completely and the greater the impoundment the more useful the mill.

Decline of Atlantic salmon

A glacial map of Europe.

The only route for a fish around a Roman mill would have been the relief channel – usually the old river – but even these were controlled with gates. On the Wycombe Stream in Buckinghamshire, a tributary of the Thames that still holds trout but has no salmon, there were 20 Domesday mills on a river only a dozen miles long. The Mole, another Thames tributary, also had 20 mills. On the Colne there were 33. Most rivers in lowland Britain, certainly most of the tributaries, were actually at mill-saturation – the point where you simply can’t fit any more in – by the year 1086. When it comes to the salmon of lowland Britain, it’s not that they were never there, but rather that we shut them out over a thousand years ago.

With salmon it has been one extinction after the next. Scroll back a few thousand years and the deforestation enacted by neolithic farmers is still discernible in our floodplains – the landscape literally rose with the silt loading in our rivers. Britain’s covering of native woodland was cut in half before the year 2,500BC. The impact on salmon would not have been slight: rivers warmed, flow regimes became more violent. The impact of the ox-drawn medieval ploughshare is also there in the stratified layers of alluvial deposits along the riverbanks of England.

And if the mills hadn’t already blocked the spawning streams of the south and east, a surge in the development of navigable waterways over the centuries of the Middle Ages will have driven the last few nails into the lowland salmon’s coffin.


Then the Industrial Revolution brought with it not only more mills, bigger, more imposing and sturdy enough now to assail the more energetic rivers of the uplands, the north and west, but another barrier to salmon migration in the form of sewage and industrial waste.

The River Mersey is just one example of a formerly clear-watered salmon stream that by 1850 was biologically dead, because of mills, coal, geology and careless man. A bewildering three-quarters of a million tons of ash, colliery waste, rubbish and sewage was tipped into the Mersey every year in the middle of the 19th century. So much that it accumulated into islands in the stream, islands wreathed by a poisonous, fetid liquor in which nothing could live, let alone the salmon.

Decline of Atlantic salmon

An engraving in Punch, casting pollution as the sceptre of death.

The Mersey was just one of many. By 1877 Archibald Young assembled a list of British salmon rivers seriously impacted by industrial pollution, mining, sewage, dams and weirs: the rivers Axe, Camel, Dart, Dee, Dove, Eden, Exe, Fowey, Kent, Ogmore, Rhymney, Ribble, Severn, Stour, Tamar, Taw, Torridge, Tees, Teifi, Teign, Towy, Trent, Tyne, Usk, Wear and Wye were all on it. The Mersey, Don, Aire and Calder were devoid of salmon. As, of course, was the Thames.

In the war to save the king of fish we are mostly losing ground. Even if salmon are now returning to some post-industrial rivers in the north, the indigenous salmon of East Anglia, of the Midlands and the south-east are long gone. While on the south coast a half-dozen chalk-fed rivers in Hampshire, Wiltshire and Dorset hold the last lowland native-salmon runs in the country, teetering on the edge of survival.

In lowland Britain, only the Thames carries a hope of what could come again. There are a few salmon in the Thames but they are not descendants of the infamous 1980s restocking programme. Rather, they are strays from the River Itchen’s wild stock, showing that if the water is clean enough and the physical or chemical barriers are taken down, then salmon can and will find their way back, as they have most famously into the Tyne and the Taff, and most heart-warmingly into the Mersey and the Aire and the Calder and the Don.


It is an exciting prospect to think that rivers from which the salmon has been shut off may recolonise in this way. The salmon returns to its natal river with almost unerring accuracy, but a few do go astray – and thus they colonise new places or, more vitally, recolonise places once lost.

And, of course, it has happened before, many times over. Since its evolution as a species a few million years ago the Atlantic salmon has been shut off time and again by glaciers and has recolonised that ground when conditions allowed. Across the geological timescales that have tick-tocked the ebb and flow of ice-sheets and indigenous races of salmon distinct to given rivers, the 1,000 years that have elapsed since mankind started to block salmon runs with water-mills is put into some kind of perspective.

The Devensian glacial epoch occurred some 17,000 to 71,000 years ago, and the Anglian 424,000 to 478,000 years ago. Before and between those major glaciations there is evidence of many other phases of glaciation, of the ice sheets breathing in and out over the European landmass and across the centuries.

Decline of Atlantic salmon

Constable’s Parham Mill, Gillingham (1826) – but mills blocked salmons’ migratory routes.

The Atlantic salmon survived them all.

With so much of the Earth’s water locked up in ice, the sea retreated beyond the tips of Britanny, Cornwall and Western Ireland. The North Sea was not a sea at all but a vast land-based ice-field frozen solid as far south as Lowestoft. This wall of ice blocked the northwards flowing alignments of the Rhine, Seine and Thames, which flowed instead into a vast glacial lake, the overspill and eventual destruction of which cut the straits of Dover and made Britain an island.

Our present-day south coast was hundreds of miles inland and all of these rivers along it, from the Thames to the tip of Cornwall, were periglacial or tundra headwaters of a vast catchment that included the Rhine and Meuse, the Seine and the rivers of Belgium, the Somme and northern France, too. This catchment marked the northern edge of the salmon’s range and its quite probably innumerable salmon migrated to and from an estuary now lost under the waves somewhere between Penzance and Brest.

Which amazing thought brings us to a final, even more amazing thought: the few salmon hanging on in that outpost of Wessex chalk have been there since the Atlantic salmon itself evolved as a species.

The southern edge of the British Isles has never been covered by ice. The blowsy rivers Test and Itchen would once have looked like the treeless spring creeks of Iceland, where trout and salmon thrive in the constant temperature of groundwater flows within miles of glacial icecaps. The race of salmon I watched running the Stour, when I left university and went to teach in Dorset in 1987, the last time I saw them there, had been running the River Stour for millions of years, not thousands.

It would seem to me that in a warming world, the repository of cool-flowing groundwater rivers across the breadth of southern England, the rivers we shut salmon out of before William invaded in 1066, might yet offer something of an arc to our oldest native vertebrate. Isn’t it about time we let the poor fish back in?