Ecological succession
A Socratic walk-through of ecological succession — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does a burnt hillside come back through plants that cannot survive in the forest they end up creating?
Burn a hillside and the first things back are fireweed, bracken, aspen, birch — and none will be there in a century. Worse, the forest that replaces them is one their own seedlings cannot grow in. It looks like a relay in which each runner hands on the baton and then dies, which invites a tidy story: each stage prepares the ground for the next, and the community climbs toward a settled climax. That story is old, influential and mostly wrong. Why is the pattern real when its usual explanation is not?
Reasoning it through
REASONING #Begin with the pioneers and ask what kind of plant wins an empty, sunlit, ash-covered hillside. It has to arrive — tiny seeds, in enormous numbers, blown far, or already lying dormant in the soil, or held in cones that open only in fire. It has to grow immediately in full sun, so cheap thin tissue and fast height. It need not last, because the site will not stay empty.
Now ask what wins in a closed forest a century later. Almost nothing arrives there; the contest is decided under shade, among seedlings. What is needed is a large, heavily provisioned seed that can sit in deep shade for years on its parent's savings, and dense, defended, slow tissue.
Those two lists are opposites, and for a hard reason rather than a stylistic one. A fixed budget for seed can be split into many small seeds or few large ones, not both. Tissue can be cheap and fast or dense and durable, not both. No plant holds both ends of these trade-offs, so the same site is won by different species at different moments. That is the whole engine, and it requires nobody to prepare anything for anyone.
Now the second half of the question, which is where the folk story is most misleading. Do the pioneers die because they built a forest they cannot live in? Mostly, no. An aspen or a birch that got established in the open can stand in a closing forest for decades. What fails is not the adult, it is the next generation: their seedlings need light and there is none. Succession is overwhelmingly a recruitment failure, not a mortality event. The pioneers are not killed by their creation; they simply stop being replaced.
Does facilitation — the genuine preparing of ground — ever happen? Yes, and it matters most where the substrate is truly bare. On a glacier foreland or fresh lava there is no nitrogen at all, and nitrogen-fixing plants such as alder or Dryas really do make the site habitable for what follows. But even the textbook case is shakier than it looks: when the classic Glacier Bay sequence was checked against actual stand histories rather than inferred by comparing different-aged sites, several stands turned out to have started with communities the single-trajectory story did not predict. Reading space as if it were time is convenient, and lossy.
Connell and Slatyer set out the alternatives in 1977 and they are still the right frame: early species may facilitate later ones, may simply be tolerated alongside them, or may actively inhibit them and hold the site until something kills them. Inhibition is common — a dense bracken or grass sward can suppress tree seedlings for decades, and the trajectory resumes only when a gap opens.
Which brings the sharpest correction: there is no climax. Disturbance recurs at intervals often shorter than the trajectory needs, so a landscape is a mosaic of patches at different stages rather than a single community approaching a destination. And some trajectories never return. Burn the same hillside again before the trees have set seed and it can lock into a grass-fire cycle: grass carries fire more readily, fire kills tree seedlings, and the feedback holds the site out of forest indefinitely. That is a genuine alternative stable state, not a delayed stage.
Where this sits next to its neighbours: the collection's piece on weed seed banks describes the same buried store of dormant seed, but reaches the opposite fixed point — there, disturbance repeats often enough that the site never leaves stage one, whereas here a single disturbance lets the trajectory run forward. The frequency of disturbance, not the store, decides which you get. And the piece on the forest height race concerns competition for light within a stand at one moment; this one is about which species can recruit at all, across stands over time.
The analogy
THE ANALOGY #Think of a building site cleared by a fire. The first trades on it are those who can be there tomorrow with a van — fencing, scaffolding, temporary power. They are not preparing the site out of goodwill; it simply favours whoever shows up fastest and can work in the open. The finishing trades come later not because the scaffolders invited them but because the job has changed into one that rewards patience and expensive equipment, and by then nobody is hiring scaffolders.
There is no client, no schedule and no job — nothing is coordinating the sequence, and the trades are not called in but arrive by seed and by luck. And a real site finishes; a hillside never does, because another fire will re-clear it before any final state is reached.
Clarifying the model
THE MODEL #Three refinements. First, the sequence is a consequence of trade-offs among independently arriving species, not a programme the community runs — which is why any given hillside's composition is partly a matter of what was nearby and what year it burned. Second, "pioneers cannot survive in the forest" should be read as "pioneers cannot regenerate in the forest"; that is why old birch stand in mature woods for decades. Third, the pioneers are usually not colonists at all in the sense of arriving from far away — much of the first flush comes from seed already in the soil, resprouting roots that survived the fire, and cones that needed the heat to open. The site was preloaded.
A picture of it
THE PICTURE #How to readThis is a version-history diagram repurposed as a map of trajectories, so treat each dot as the site's condition rather than a code change, and read the two rows as alternative futures from the same fire — horizontal position counts steps along each line, not a shared calendar. The line running through to Forest again is the familiar return: the preloaded seed bank, then fast light-demanding trees, then a canopy that closes and admits only shade-tolerant seedlings. The other line splits at Fire and never rejoins — a second burn before seed is set, then grass, then a feedback that holds the site out of forest. The point of drawing it as a branch is that nothing forces the two lines back together.
What became clearer
WHAT CLEARED #Succession is not a relay and not a climb toward a destination. It is the same site won by different specialists as its conditions change, because the traits that let a plant seize open ground are the physical opposite of those that let one recruit under shade. The pioneers do not sacrifice themselves; their adults often persist and only their offspring are shut out. And the sequence is a tendency, not a law — disturbance restarts it, inhibition stalls it, and a feedback can lock a hillside into a state it never leaves.
Where to go next
ONWARD #- Why some ecosystems, such as many grasslands and heaths, are maintained by the disturbance that would reset a forest.
Key terms
TERMS #| Term | What it means |
|---|---|
| Pioneer species | a fast-growing, light-demanding, widely dispersed species that establishes on newly opened ground. |
| Facilitation, tolerance, inhibition | Connell and Slatyer's three ways early occupants can affect later ones: helping them, neither helping nor hindering, or actively suppressing them. |
| Alternative stable state | a persistent condition, held by its own feedback, that a site can enter and not leave without a further shock. |
Every term the collection defines is gathered in the glossary.