Salmon fly fishing

Fishing an artificial fly for Atlantic salmon and sea trout in rivers, presented down and across the current on a tight line so that the flow swings it over the fish, usually with a double-handed rod. Distinct from dry fly fishing, which presents a floating fly to a feeding fish, and from stillwater fly fishing, which has no current to do the work.

It is how Atlantic salmon are fished for almost everywhere they are fished for, and on many rivers it is the only method permitted for part or all of the season. It searches large water systematically with one rod, it works at a range and a depth no other freshwater fly method reaches, and it is the technique the whole culture of beats, ghillies and named pools is organised around.

THE CURRENT IS THE ENGINE. The fly is cast across and down, the line is mended upstream or downstream to slow or speed the swing, and the fly fishes on a tight line all the way round until it hangs below the angler. Where the fly is in the water column and how fast it is crossing are the two variables under control, and they are controlled by line density, by the angle of the cast and by mending rather than by retrieving.

THE FLY IS CHOSEN BY WATER, NOT BY HATCH. Height, clarity and above all temperature decide size and line. The traditional rule puts the change at about 48-50°F: colder than that, a larger fly fished deep on a sunk or sink-tip line; warmer, a smaller fly close to the surface on a floater. Published fly-size guidance follows the same curve, running from roughly size 2-6 in water in the mid-forties down to size 10-12 as it passes sixty.

THAT THRESHOLD IS CONTESTED AND IS RECORDED AS CONTESTED. Practitioners argue the 48-50°F line owes as much to the popularity of greased-line literature as to the fish, and that a sinking line is the right answer at any temperature where the flow needs countering to hold the fly in the taking zone. The heuristic is genuinely useful and is genuinely not a law.

THE STEP-DOWN IS THE SEARCH PATTERN. Cast, fish the swing out, take a pace or two down, cast again. It is what makes a long pool a finite problem.

AND THE GHILLIE OR BOATMAN IS PART OF THE METHOD, NOT A LUXURY. On big let water the person who knows where the fish lie, what the height is doing and which line to put on is doing a job the angler cannot do from a single visit, and on rivers fished from a boat they are also handling the boat over the lie.

Suits rivers with enough flow to swing a fly, and works through the range of heights between a river too low to hold running fish and one too coloured to fish. Cold water favours a bigger fly fished deeper, warm water a smaller fly near the surface. Ruled out by a river in heavy spate, by ice, and by any period when the season is closed or the beat is fly-only and the angler is not equipped for it. Conditions decide this method more than any other form of freshwater fly fishing.

An Atlantic salmon stops feeding when it enters fresh water and does not start again before it spawns. It lives on what it built at sea, and every fish an angler covers on a river is one that will not eat the fly.

SO WHAT IS BEING PROVOKED IS NOT HUNGER. The explanations offered are territorial irritation, a reflex left over from the feeding life at sea, or simple aggression, and no one of them is settled. What matters practically is that the fly is not competing with a food supply and does not have to look like one — which is why a sport with a reputation for fussiness tolerates patterns that resemble nothing that has ever lived.

AND IT REORGANISES EVERY DECISION. Because there is no hatch to match, the fly is chosen from water temperature, height and clarity. Because the fish is not searching for food, it must be shown the fly at its own depth and at a speed it will react to, which makes line density and mending the real tackle decisions. Because a fish that will not eat cannot be tempted twice in the same way, the pool is covered systematically rather than worked over.

IT ALSO EXPLAINS THE CATCH RATE. Long blank periods are the ordinary experience of this method rather than evidence of doing it wrong, and any account of salmon fishing that implies otherwise is selling something.

Salmon fishing is among the most closely regulated angling there is, and the regulation is set river by river and revised year by year.

THE SEASON IS LOCAL. Close times differ between rivers and between jurisdictions, and a date that is correct on one river is wrong on the next one along the coast. Many rivers additionally run fly-only periods at the start and end of the season, and some restrict methods further — the prohibition of treble hooks on particular beats being a common example.

THE CONSERVATION MEASURES ARE RESET ANNUALLY. Scotland grades its salmon rivers by statutory instrument each year, with only the poorest grade carrying mandatory catch and release as law, and the grading vocabulary itself changed before the 2025 season from numbered grades to good, moderate and poor conservation status. A river's grading is therefore both a moving number and, in older sources, a moving language.

AND SOME RIVERS ARE THEIR OWN JURISDICTION. The Tweed district is regulated under its own legislation across a catchment that includes English ground, and no Environment Agency rod licence is required anywhere in it. The Border Esk runs the other way: mostly in Scotland, administered by the Environment Agency, and an English licence is required. Answering either from which country the bank is in produces the wrong answer.

FISHINGHQ PUBLISHES NO DATE, GRADING OR KILLING LIMIT. It has no mechanism for following an annual instrument, and a figure reproduced from a document read once looks exactly as authoritative as a correct one while being wrong. The river's own governing body — a district salmon fishery board, a river commission, or the national authority — is who to ask, and the beat will know what applies on the day.