Skip to content
NatureHQ

A guided journey · 4 stops

The long way home

Salmon do smell their way home. They also cross an ocean where there is nothing to smell — and the two halves of that journey needed two different answers, sixty years apart.

The route

  1. Smell, and how far it really works
  2. Salmon
  3. Magnetoreception
  4. Sea turtles

"Salmon smell their way home" is the rare popular claim that is genuinely true, decisively demonstrated, and still only half the story. Following it properly is a good way to see how a scientific answer gets built, corrected and extended without ever being overturned.

The route starts with the sense, because the strength of the salmon result depends on understanding what smell can and cannot do — it works on a plume, at close range, and a river's odour is undetectable in the open Pacific. Then the experiment that settled the river half, which is one of the most elegant in biology: rather than trying to prove fish were following their home water, somebody gave them a smell that exists in no water at all.

Then the part the popular version leaves out, and the animal that supplies the reason to take it seriously.

  1. Stop 1 of 4. Senses and abilities

    What smell can actually do

    Start here and the salmon story stops being magic. A plume is not a cone — turbulence tears it into filaments, and tracking one is a navigation problem rather than a detection problem. That framing is what makes the scale limit obvious later: a river mouth is a source, and the open ocean is not.

    Read Smell, and how far it really worksA moth can detect a single molecule and still cannot fly straight to the source. The nose was never the limiting factor.

  2. Stop 2 of 4. Marine life

    The experiment that settled the river half

    The morpholine study is the payoff: expose juveniles to a compound found in no stream anywhere, and eighteen months later the adults arrive at whichever stream you dosed with it. A fish returning to morpholine is not returning to a place. It is returning to a memory, and that is what a substitution experiment can show and an observation never could.

    Read SalmonSalmon do smell their way home — for the last few kilometres. A river’s scent does not reach the open Pacific, and something else brings them to the right coast.

  3. Stop 3 of 4. Senses and abilities

    And the half a nose cannot explain

    Once the odour answer is properly bounded, the gap is obvious and large: thousands of kilometres of ocean where no river scent reaches. This page supplies the candidate — and, just as importantly, the reason to hold it loosely, since the receptor that would make it work has never been found in any vertebrate.

    Read MagnetoreceptionA bird’s magnetic compass has been demonstrated for fifty years, in laboratories all over the world. Nobody has found the receptor.

  4. Stop 4 of 4. Reptiles and amphibians

    Why anyone believes the ocean half at all

    The salmon evidence for magnetic imprinting is correlational, and correlations across decades of catch data deserve scepticism. The reason it is taken seriously is that the same imprinting logic has direct experimental support in an unrelated animal that solves the same problem — which is what this last step supplies, and why the route ends somewhere other than a fish.

    Read Sea turtlesA turtle does not return to a beach. It returns to a magnetic signature — and when the signature drifts along the coast, the nests drift with it.

Where this leaves you

You end with a two-scale answer and an honest difference in confidence between its halves. The river choice is settled by experiment: fish exposed to a synthetic compound as juveniles returned to it eighteen months later. The ocean crossing is not settled — geomagnetic imprinting is the leading candidate, and the evidence is fifty-six years of fishery records tracking the drift of the field rather than an experiment on a fish.

What the route does not settle is what a stream's odour actually consists of. The imprinting result is fifty years old and the chemistry it depends on has never been characterised, which means nobody can say how stable a home smell is across years of changing flow and land use.

Last reviewed 2026-09-03. All guided journeys