The Shape of Stasis

A guitar string, tuned and at rest, looks like it’s doing nothing. It’s straight, still, boring. But it is not relaxed — it is under tens of kilograms of tension, held taut between two anchors. Pluck it and it snaps back to the same straight line, every time, because forces are actively pulling it there. The stillness is the achievement, not the absence of one.

That is the single hardest idea in this whole course, so we spend a lesson on it: a flat line is not an empty waiting room. When a species sits morphologically unchanged for millions of years, something is holding it still — and holding hard. This lesson makes the shape of that stillness visible, then names the forces doing the holding.

Change is lumpy, not smooth — the shape

Draw the gradualist’s picture of change and you get a ramp: a straight line tilting steadily upward, a hair of change every generation, forever. Under this view — phyletic gradualism — a species is always slowly becoming its descendant, and the fossil record should read like a smooth gradient.

Now draw what the rocks actually show and you get a staircase: long flat treads (stasis) snapped by short, near-vertical risers (punctuation). This is the shape at the heart of punctuated equilibrium, the model Niles Eldredge and Stephen Jay Gould proposed in 1972: most species spend most of their existence changing very little, interrupted by geologically brief bursts of change concentrated at speciation.

The two pictures are genuinely different geometries, and they predict different things:

The rest of this lesson is about the flat parts. (The next lesson takes the risers.) Because the flat parts are where the counterintuitive action is: nothing visibly moves, yet the system is working hard.

Drive the shape yourself

Below is the course’s stasis-and-lurch lab. The main panel tracks one lineage’s trait across a long horizon; the inset shows the same story as a fitness landscape — a population sitting on a peak, held there until a shock lets it cross a valley to a higher peak.

Three things to notice as you play:

• The flat stasis spans dominate. Even at a middling constraint, the line spends most of its length dead flat. Change is the exception, not the baseline — exactly backwards from the ramp.
• The lurches are near-vertical. When change comes, it comes fast and steep, then freezes at a new level. There is almost no diagonal drift — no slow ramp between plateaus.
• The inset population is pinned. Watch the ball sit on its peak through the whole flat span. It only crosses the valley to the new peak at a lurch — and only then does the “new peak” label even appear.

Crank the constraint to the top and shocks to the bottom, and the lineage barely moves for the entire horizon. That extreme — near-total stasis — is not a broken simulation. It is a horseshoe crab.

Stasis is active, not idle — the four mechanisms

If a form can sit unchanged for millions of years, something is enforcing that. Punctuated equilibrium leans on four forces, working together. None of them is “selection switched off.”

1. Stabilising selection

The mechanism. Of the three flavours of selection, this is the conservative one. Directional selection pushes a trait one way (bigger, faster); disruptive selection splits it toward two extremes; stabilising selection culls both extremes and holds the mean in place. Picture the bell curve of a trait — say, beak depth. Stabilising selection shaves off the too-big and too-small tails every generation, so the average stays parked on the optimum.

Worked example. Human birth weight is the textbook case. Babies much lighter than ~3.5 kg face higher mortality; babies much heavier face delivery complications. For decades, mortality data traced a clean U-shape with its floor right at the population mean. Selection there isn’t building a better baby — it’s a wall on both sides holding the average where it is.

The pitfall. “No change” tempts you to conclude “no selection.” It’s the opposite: a form staying flat under a full spread of mutations each generation is evidence of ongoing, vigorous selection. Silence here is the sound of a lot of pruning.

When it dominates: a well-matched form in an environment that isn’t moving. The optimum is where you already are, so every step off it is downhill.

2. Developmental and genetic constraints (canalisation)

The mechanism. Some changes are hard to make, not just hard to keep. A body plan is a deeply interlocked developmental program: alter one early step and a cascade of downstream structures breaks. Canalisation (C. H. Waddington’s term) is development’s tendency to buffer itself, funnelling a wide range of genetic and environmental inputs toward the same reliable outcome — like a marble rolling down a canal that keeps re-centring it. Deviations get absorbed before they ever reach the phenotype.

Worked example. The tetrapod limb has settled on one-bone-then-two-bones (humerus, then radius + ulna) for ~375 million years, across frogs, bats, whales, and you. Not because five-bones-then-one couldn’t in principle be useful, but because the developmental machinery that lays down limbs is so entangled that gross rearrangements almost always produce a non-viable embryo. The blueprint resists its own editing.

The pitfall. Don’t read a constraint as “selection prefers this.” Sometimes the alternative is simply developmentally unreachable — never offered to selection at all. Absence of variation is not the same as selection against it.

When it dominates: deep, ancient, load-bearing features — body plans, core biochemistry — where the surrounding system has grown up assuming they hold.

3. Gene flow across a large population

The mechanism. A big, well-mixed, interbreeding population is a change-averaging machine. Suppose a local subgroup starts drifting toward a novel form. Migrants from the rest of the range keep arriving and interbreeding, diluting the local variant back toward the species average every generation. Gene flow — the movement of alleles between groups — swamps local innovation. The bigger and more connected the population, the harder it is for any corner of it to go its own way.

Worked example. This is why Eldredge and Gould (following Ernst Mayr) expect new forms to arise in small, peripheral isolates, not in the big central population. A pond cut off from the river can shift fast; the river itself, constantly mixing, stays put. Hold this thought — it’s the hinge of the whole model, and the engine of the next lesson.

The pitfall. Intuition says a bigger population evolves faster (more mutations, more raw material). For directional change it’s often the reverse: large size and high connectivity are stabilising, because mixing averages novelty away faster than it can accumulate.

When it dominates: widespread, mobile, densely connected species with no strong barriers inside their range.

4. “If it ain’t broke” — the plain logic

The mechanism. Strip away the biology and a simple economic truth remains: when a form already works, the expected payoff of a large change is negative. Most mutations of large effect are harmful; the current design has been road-tested for millions of years; the environment isn’t demanding anything new. So the safe bet — the one selection keeps making — is stay.

Worked example. The shark’s basic body plan is ~450 million years old. Sharks aren’t “unevolved”; they are a solution so good that the ocean has repeatedly failed to offer a better nearby one. Stasis here is not laziness. It is a form that keeps winning.

The pitfall. “If it ain’t broke, don’t fix it” quietly assumes the environment stays still. Stasis is only optimal until the landscape moves — at which point the very forces that held you flat become the reason you can’t keep up. (Lesson 2’s problem, and lesson 4’s corporate graveyard.)

A worked case with real numbers: the trilobite eye

Punctuated equilibrium didn’t arrive as pure theory — it came out of Eldredge’s dissertation on a Devonian trilobite, Phacops rana. Trilobites had schizochroal compound eyes built from vertical columns of lenses, and the count of those lens columns is a clean, countable, fossilisable trait — a rare gift for measuring change.

Across roughly eight million years of Middle Devonian rock (about 390–380 million years ago), Eldredge tracked the lens-column number through many populations and long spans of section. The result that founded the model: for most of that time the count sat flat — stasis, holding steady across immense stretches. The change that did occur wasn’t a smooth ramp threading the whole record; it was a step, a reduction (commonly cited from 18 columns to 17) that showed up abruptly, appearing first in a peripheral population and then spreading — not a gradual slide averaged over the eight million years.

Put the timescale in human terms. Eight million years is roughly 400,000 human generations. The gradualist ramp predicts a visible tilt over a span like that. What the rock delivered was a long horizontal line with a step in it — the staircase, measured lens by lens.

For comparison, three poster children of extreme stasis, with their timescales made explicit:

None of these is a lineage that “forgot” to evolve. Each is a form so well-anchored that the four brakes held for a geological age.

Watching near-total stasis

Now push the lab to an extreme: a heavily locked-in lineage in a mostly calm world. This is the horseshoe-crab regime.

Notice what the constraint dial is really doing: it’s raising the threshold a shock must clear to move the trait at all. A locked-in lineage isn’t one that nothing happens to — it’s one where most of what happens gets absorbed. The world keeps knocking; the door holds.

When to reach for stasis-thinking

Stasis-thinking is the move of asking, “what is actively holding this still?” — and it’s the right tool in a specific situation.

• Reach for it when a system has looked unchanged for a suspiciously long time: a market with the same three players for a decade, a codebase no one refactors, a body plan 400 million years old, a personal habit that survives every New Year. The stability is a fact to be explained, and the explanation is a set of forces you can name — and, later, break.
• The payoff: naming the brakes (which stabilising force? which lock-in? what’s being averaged away?) tells you where the pressure is stored and what would have to give for change to come. That’s a far better map than “it’s just how things are.”
• The trade-off / where it misleads: don’t over-apply it to genuinely fast-moving systems, where change really is the baseline — forcing a stasis frame there invents brakes that aren’t holding. And beware the flip side: stasis-thinking can lull you into treating a flat line as permanent when it’s actually stored pressure about to release. The frame’s job is to make you ask when, not to promise never.

Check yourself

Recap

Where this goes next

We’ve spent this lesson on the treads of the staircase — the flat parts, and the four brakes that hold them. But a coiled spring exists to be released. Lesson 2 takes the risers: the punctuation and its triggers. What breaks a stasis that four forces are actively defending? The answer is the hinge we flagged under gene flow — a small peripheral isolate slips free of the averaging, a shock removes a constraint, and a threshold tips (hello, critical mass). The stored pressure discharges, the trait lurches, and a new flat line begins. Bring the two questions with you: you now know what holds the door shut, so next you’ll learn exactly what kicks it open.