Punctuated Equilibrium

How punctuated equilibrium works

Pre-1972 palaeontology textbooks drew fossil lineages as smooth diagonal lines — body size creeping upward through millions of years, the way Darwin imagined gradual transformation. The trouble was, when palaeontologists actually measured fossils through deep rock sequences they kept finding the opposite: long stretches where nothing changed, broken by sudden jumps where new forms appeared as if from nowhere. The standard explanation was that the rock record is too gappy. Eldredge and Gould looked at the same data and proposed that the rock record is showing the truth — evolution does work in jumps.

The model has two parts. Stasis: once a species is well-adapted to its niche, stabilising selection and gene flow across its range hold morphology steady for millions of years. Lingula brachiopods have looked essentially the same for 450 million years; horseshoe crabs for 150 million; coelacanths for at least 80. Punctuation: when change does happen, it happens in small peripheral populations during peripatric speciation, takes a geologically short time (5,000–50,000 years), and the new species then displaces or coexists with the old. The transition is so brief that it falls inside a single sediment layer.

Drawn as morphology against time, the pattern looks like a stretched comb — long horizontal stems for stasis, near-vertical risers for punctuations. Phyletic gradualism would predict a smooth diagonal staircase. The difference is empirically testable, and palaeontologists have run the test many times since 1972.

Why species stay still

The hardest part of the model for many biologists was not the punctuation but the stasis. If selection prunes unfit variants and mutation never stops, why would morphology not drift continuously? Three reasons, all empirically supported:

• Stabilising selection. A well-adapted species sits in an adaptive peak; deviations are pruned. Bumpus's sparrows after the 1898 Providence storm are the micro-example: extreme individuals died, the average survived.
• Gene flow as homogenizer. Across a wide range, dispersal mixes alleles between populations; local drift cannot accumulate because new arrivals dilute it.
• Genetic homeostasis. Coadapted gene complexes resist disruption. Pulling one allele off-target breaks downstream regulation. Phenotype is canalised in Waddington's sense.

Hunt (2007) compiled 250 fossil time series across phyla and found stasis was the most common mode in two-thirds of cases. Random walks were second. Smooth directional change — phyletic gradualism — was a minority pattern.

Punctuated equilibrium vs phyletic gradualism

	Phyletic gradualism	Punctuated equilibrium
Tempo of morphological change	Slow, steady, continuous	Concentrated in short bursts
Distribution over a lineage's lifespan	Spread evenly	~99% stasis, ~1% punctuation
Where change happens	Throughout entire species range	In small peripheral isolates
Speciation mode	Anagenesis (whole lineage transforms)	Cladogenesis (lineage splits)
Predicted fossil pattern	Smooth gradient through strata	Stasis + sudden replacement
Treatment of fossil "gaps"	Real preservation failures	Real biology — peripheral isolates rarely preserved
Mechanism	Continuous directional selection on whole population	Peripatric speciation + stabilising selection
Best documented in	Some foraminifera, some ammonites	Bryozoans, Cerion snails, Turkana mollusks, trilobites

The pluralist consensus in modern palaeobiology is that both patterns occur, in different lineages and even in different parts of the same lineage's history. The original 1972 paper has held up better than most evolutionary syntheses of its era; the strong claim that punctuation is the dominant mode for most groups is supported, while the strongest version (gradualism is wrong everywhere) has been softened by data.

Worked example — Cerion snails and Caribbean bryozoans

Stephen Jay Gould spent decades collecting Cerion land snails on the Bahamas, where shells preserve well in calcareous dunes and small cays act as natural founder-population labs. Across thousands of specimens spanning hundreds of thousands of years, Gould found each main island carried a conservative form varying within tight bounds for tens of thousands of years, while peripheral cays produced sharply distinct morphs that, when their populations expanded, replaced the parental form on the main island over a few thousand years. Stasis, punctuation, stasis again — peripheral cays as the laboratory.

Alan Cheetham's study of Caribbean cheilostome bryozoans (1986) is even crisper because zooid shapes have many discrete landmarks and the sediment record is dense. Across 15 million years and 17 species in the genus Metrarabdotos, most species changed almost nothing across their existence, and morphological differences between species were established within a geologically short interval — typically less than 100,000 years, often less than 10,000. Gradualism predicted a smear of intermediates; the data showed sharp jumps separated by long flat plateaus. Still cited as the strongest single test of the model.

Peter Williamson's 1981 Lake Turkana mollusks: a 5-million-year sequence shows long stasis interrupted by short bursts at moments of major lake-level change. Critics argued ecophenotypic variation; later genetic work supports the speciation interpretation for most events.

Real-world cases of stasis and punctuation

• Bahamian Cerion land snails. Stasis for tens of thousands of years on main islands; rapid morphological shifts on peripheral cays. The clearest small-organism, short-timescale example.
• Caribbean bryozoans (Metrarabdotos). 15 million years of fossil record show clean punctuations and long stasis across 17 species.
• Lake Turkana mollusks. Williamson's 5-million-year sequence; pulses of morphological change tied to lake-level shifts.
• Devonian trilobites. Eldredge's thesis lineage Phacops rana; eye facet number shows stasis interrupted by sudden shifts.
• Lingula brachiopods. Externally almost unchanged in 450 million years — the textbook poster organism for stasis.
• Horseshoe crabs. Roughly 150 million years of conserved morphology; modern species are visually indistinguishable from Mesozoic forms.
• Coelacanths. The lineage thought extinct since the Cretaceous turned up alive in 1938; its fossil and modern forms are nearly identical despite ~80 million years of separation.
• Hominin braincase volume. Punctuations at Homo erectus emergence (~1.8 Mya) and at Homo sapiens emergence (~300 kya), separated by stretches of relative stasis.

Variants and refinements

• Strong vs weak punctuation. The strong form claims essentially all change happens at speciation events. The weaker, modern form claims only that change is concentrated in short bursts; the bursts may or may not coincide with speciation.
• Species selection. Gould later proposed that, given long stasis, evolution above the species level operates by differential origination and extinction of species rather than within-species change. The clade-level analogue of natural selection.
• Coordinated stasis. Brett and Baird (1995) noted that whole faunas can stay morphologically stable together for millions of years, then shift in concert during environmental upheavals — community-level punctuation.
• Quantum evolution. Simpson (1944) had earlier proposed something close to punctuation under a different name, framed in terms of adaptive zone shifts; Eldredge and Gould give him explicit credit.
• Punctuated gradualism. Malmgren et al. (1983) for foraminiferal lineages: a hybrid pattern with long gradual phases punctuated by occasional rapid bursts. Most lineages probably look like this.

Common pitfalls

• "Punctuated equilibrium overturns Darwinism." Gould was happy to have it described as a friendly amendment, not a revolution. The mechanism is selection plus drift in small populations — exactly Darwin's ingredients, applied with Mayr's peripatric framing.
• "Bursts are saltational." They are not single-generation jumps. A 50,000-year burst is still many thousands of generations of conventional selection. Goldschmidt's "hopeful monsters" are not part of the model.
• "Stasis means evolution stops." Genetic evolution continues — alleles drift, neutral substitutions accumulate, behavioural adaptations refine — but visible skeletal morphology stays inside its envelope. Stasis is morphological, not molecular.
• "The fossil record is too patchy to test this." True for many groups, false for several where dense, well-dated sequences exist. Where the data are good, punctuation is often the better fit.
• "Punctuated equilibrium and gradualism are mutually exclusive." They are endpoints on a tempo spectrum. Most lineages show stretches of each. The empirical question is mixture, not which one is right.
• "Stasis needs no explanation." It does — selection should normally cause drift in some direction. The active maintenance of stasis through stabilising selection and genetic homeostasis is one of the model's main contributions.