Ecological Succession

Ecological succession describes how the species composition of a community changes over time at a given site, following either the creation of new substrate or the disturbance of an existing community. Ecologists distinguish two broad classes by starting condition. Primary succession begins on substrate that has never supported life — volcanic rock, a glacier's exposed till, a sand dune, a lake bed drained of water — meaning there is no soil, no seed bank, and no organic matter to start with. Secondary succession begins on a site that already has soil and often a residual seed bank or root system, because the disturbance killed the standing vegetation but left the substrate largely intact — an abandoned farm field, a burned forest, a hurricane-flattened woodland. The sequence of communities that occupies a site over successional time is called a sere, and each identifiable stage within it is a seral stage (or successional stage). The first organisms to establish on bare substrate are pioneer species — typically fast-growing, wind- or animal-dispersed, tolerant of nutrient-poor and physically harsh conditions, and short-lived. Over time these give way to species with different life-history strategies: slower-growing, longer-lived, more shade-tolerant or nutrient-demanding organisms that could not have established on the original bare substrate but thrive once earlier colonists have modified it. If succession proceeds long enough without further disturbance, the community may approach a relatively stable, self-perpetuating state historically termed the climax community — though, as this article details, the reality and universality of a single climax endpoint is one of the most productive controversies in community ecology. Three conceptually distinct mechanisms, formalized by Joseph Connell and Ralph Slatyer in an influential 1977 synthesis, explain why one set of species replaces another rather than simply coexisting indefinitely: facilitation, tolerance, and inhibition. These are not mutually exclusive descriptions of nature so much as three testable hypotheses about the causal relationship between an early colonist and the species that follow it, and different successional sequences — even different stages of the same sequence — can be governed by different mechanisms.

Step 1 — Colonization of open space. A disturbance or new-substrate event removes competitors and creates open resources (light, space, nutrients, water). In primary succession this space is essentially a blank physical and chemical slate: no organic nitrogen, poor water retention, no mycorrhizal fungi, often unstable or oxidizing mineral surfaces. In secondary succession the "blank slate" is far less blank — a seed bank, dormant rootstocks, mycorrhizal networks, and organic-rich soil usually persist, which is the single biggest reason secondary succession runs faster than primary succession.

Step 2 — Differential dispersal and establishment (why pioneers arrive first). Pioneer species share convergent traits regardless of taxon: high seed output, long-distance dispersal (wind, birds), rapid germination, tolerance of high light and nutrient stress, and often nitrogen-fixing symbioses (lichens, alder, lupine). These traits let them reach and survive on substrate that would kill most later-successional species outright. Lichens, for instance, can extract minerals directly from bare rock and fix atmospheric nitrogen via cyanobacterial partners, building the first organic film on substrate with zero prior soil.

Step 3 — Environmental modification. As pioneers grow, reproduce, and die, they change the local environment: roots crack rock and trap wind-blown dust into proto-soil; leaf litter and dead biomass add organic matter and nitrogen; canopy cover changes light quantity and quality, humidity, and soil temperature; root systems and fungal partners increase water infiltration and cation-exchange capacity. This is the crux of the mechanism — the community physically rewrites the abiotic template of the site.

Step 4 — Differential response of later colonizers: the three models.

• Facilitation. The modified environment becomes permissive or even necessary for later species that could not have established under the original conditions — nitrogen-fixers making nitrogen available to later, nitrogen-demanding species is the textbook case (e.g., alder-fixed nitrogen enabling spruce establishment at Glacier Bay). Facilitation is most clearly demonstrated in early primary succession, where abiotic conditions are initially so extreme that essentially any biological amelioration helps subsequent colonists.

• Tolerance. Later species are not helped or harmed in a decisive way by earlier occupants; they simply have life-history traits (slower growth, greater shade tolerance, lower resource requirements at maturity) that let them establish and eventually outcompete pioneers regardless of what the pioneers did to the site. Under the tolerance model, the sequence reflects differences in species' competitive ability and stress tolerance along a resource gradient that the community itself is depleting (e.g., light) rather than active help from predecessors.

• Inhibition. Early colonists actively suppress the establishment of both other pioneers and later species — through resource pre-emption, allelopathic chemicals, or dense root/canopy occupation — and are only displaced when they die, are damaged by disturbance (herbivory, disease, physical damage), or senesce naturally, opening space that a later species then wins by superior competitive ability or chance arrival. Under inhibition, succession is not a smooth relay race but a series of competitive lockouts broken by attrition.

Step 5 — Convergence toward slower turnover. As succession proceeds, biomass accumulates, nutrient cycling becomes more closed (less loss to leaching or erosion, more internal recycling through decomposers), and species turnover slows because later-successional species tend to be longer-lived and more resistant to the modest disturbances that would have wiped out pioneers. Community respiration and production tend to approach rough balance (net ecosystem production approaches zero) in mature stages, though this equilibrium is dynamic, not static, and is punctuated by internal gap-phase disturbance (windthrow, individual tree death) even within "climax" forest.

Glacier Bay, Alaska — the classic chronosequence. Beginning with William S. Cooper's fieldwork starting in 1916 and continued by successors for over a century, ecologists exploited a natural experiment: Glacier Bay's ice has been retreating since roughly 1750, exposing a spatial gradient of sites that have been ice-free for different known lengths of time — from newly exposed till at the glacier's current edge to terrain exposed over 200 years ago at the bay's mouth. Because the underlying parent material, climate, and regional flora are essentially the same across sites, differences in vegetation across the gradient can be read as a substitute for direct long-term observation — a chronosequence standing in for a time series. Cooper documented a strikingly consistent sequence: pioneer mosses and Dryas (a nitrogen-poor-tolerant mat plant) on the youngest till, giving way within decades to nitrogen-fixing alder (Alnus) thickets, which in turn were succeeded over roughly a century by Sitka spruce, and eventually by spruce-hemlock forest on the oldest, longest ice-free terrain. Soil nitrogen measurements confirmed the mechanistic inference: nitrogen content rose sharply under alder stands and remained elevated in later spruce stands, directly supporting facilitation as the operative mechanism for this stage of primary succession. This chronosequence approach has an important limitation later ecologists were careful to flag — it assumes each site's future will mirror another site's present, which is only valid if the sites are genuinely comparable and if succession is not strongly historically contingent, an assumption not always safe to make.

Krakatoa — the pace of primary succession from true zero. After the 1883 eruption obliterated all life on the remaining islands, biologist Melchior Treub's 1886 survey found the earliest colonists: cyanobacteria crusts, ferns, and a few windblown seeds. By 1897, roughly 62 species of vascular plants were recorded; by 1919, dense forest with over 300 species existed on the tallest island. Because the starting point was unambiguously lifeless (verified by contemporaneous accounts of total incineration), Krakatoa is often cited as the cleanest real-world demonstration that a fully complex forest community can self-assemble from literal zero within a human lifetime, underscoring both the rapidity of colonization by wind- and sea-dispersed pioneers and the strong directional, non-random character of the assembly sequence.

Piedmont old-field succession — quantifying secondary succession. Extensive studies of abandoned agricultural fields in the southeastern United States (synthesized by Catherine Keever in the 1950s and subsequently many others) tracked secondary succession on soils that retained fertility and seed banks. The sequence is remarkably reproducible: crabgrass dominates within the first year, horseweed and ragweed (annual forbs) dominate year two, followed by broomsedge and asters (perennial forbs and grasses) over years 3–15, pine (loblolly or shortleaf, depending on region) establishing canopy dominance by roughly year 25–100, and shade-tolerant hardwoods (oaks, hickories) slowly replacing pine beneath its own canopy over a further century, because pine seedlings cannot establish in their own species' shade while oak seedlings can — a demonstration of the tolerance model operating at the later stage of an otherwise facilitation-flavored early sequence. Species richness in these studies typically rises through the first several decades and then plateaus or slightly declines as canopy closure reduces understory light, an early empirical hint of the diversity pattern that the Intermediate Disturbance Hypothesis later formalizes.

Connell and Slatyer's conceptual test (1977). Rather than a single experiment, Joseph Connell and Ralph Slatyer performed a synthesis of existing successional studies (rocky intertidal communities, forests, old fields) and proposed the three-model framework as explicitly falsifiable: each model predicts a different outcome if the early-successional species is experimentally removed. Under facilitation, removing the pioneer should delay or prevent the establishment of later species; under tolerance, removal should have no effect on the later species' eventual arrival or growth rate; under inhibition, removal should accelerate the arrival of later species. Subsequent removal experiments in old-field and intertidal systems (for example, clearing pioneer barnacles or algae to test whether later species colonize the cleared patches faster) generally support inhibition as common in secondary succession among competitors for the same limited early resource (space, light), while facilitation is disproportionately important on physically extreme primary-succession substrates. This turned succession from a purely descriptive pattern into a hypothesis-testing research program.

Example 1 — Reading a chronosequence to infer mechanism. A survey at Glacier Bay records soil total nitrogen (g N per m² in the top 20 cm) across sites of known deglaciation age:

Reasoning: Soil nitrogen rises roughly 24-fold between the 5-year and 80-year sites, coinciding exactly with the alder stage, then plateaus once spruce dominates and alder declines (spruce shades out alder, and spruce needle litter decomposes more slowly, adding less N per year but conserving the pool built by alder). Because nitrogen is typically the limiting nutrient in these glacial soils, and spruce is a nitrogen-demanding species that performs poorly on unamended till, the pattern is diagnostic of facilitation: alder is not merely present before spruce, it is causally enabling spruce's later success by fixing nitrogen the spruce could not otherwise obtain. A tolerance-only model predicts no systematic soil change caused by the pioneer; an inhibition model predicts alder should suppress, not assist, the following species. The data support facilitation specifically at this transition — while the pine → hardwood transition in Piedmont old fields, by contrast, is a tolerance case (pine does nothing to help oaks; oaks simply tolerate shade that pine cannot).

Example 2 — Predicting successional trajectory after two different disturbances. A temperate forest region experiences two disturbance events in the same decade: (A) a lava flow buries 40 hectares under 3 meters of fresh basalt, and (B) a wind storm flattens 40 hectares of standing forest without disturbing the soil. Reasoning: Site A must undergo primary succession: no seed bank or soil survives 3 meters of basalt, so colonization must begin with wind-dispersed lichens and spores; nitrogen-fixing pioneers will likely be essential before nutrient-demanding trees can establish; the full return to closed forest will plausibly take centuries, following the Glacier Bay timescale. Site B undergoes secondary succession: the soil, seed bank, root suckers of resprouting shrubs, and mycorrhizal networks survive the storm intact, so recovery skips the lichen/moss/bare-rock stages entirely and jumps directly to fast-growing annual and perennial herbs within the first growing season, following the Piedmont timescale of decades rather than centuries to reach forest again. The key predictive variable is not the severity of canopy loss (both sites lost 100% of standing vegetation) but whether the soil and biological legacy (seed bank, mycorrhizae, nutrient pool) survived — this is precisely the distinction that separates primary from secondary succession and explains the roughly order-of-magnitude difference in recovery time between the two sites.

Frederic Clements (1874–1945) was the dominant figure in early American plant ecology, trained at the University of Nebraska during a period when grassland ecologists were obsessively documenting vegetation change across the Great Plains. Clements developed an extraordinarily systematic — and extraordinarily influential — theory in his 1916 monograph Plant Succession: An Analysis of the Development of Vegetation, arguing that vegetation development followed the same predictable stages as an organism's growth: initiation, a sequence of developmental stages, and maturation into a stable climax determined essentially by regional climate alone, regardless of the specific disturbance or starting conditions. He treated the plant community almost literally as a "superorganism" with emergent, quasi-physiological properties. His rival, Henry Gleason (1882–1975), a botanist working largely independently, argued in his 1926 paper "The Individualistic Concept of the Plant Association" that there is no such organismic unity: each species responds independently to the physical environment according to its own tolerances, and communities are contingent, loosely bounded assemblages of species whose co-occurrence reflects overlapping individual ranges rather than a coordinated developmental program. The Clements–Gleason debate structured community ecology for the following half-century and still underlies live arguments about whether communities have emergent, predictable properties or are better understood as statistical accidents of individual species distributions — a debate this article's treatment of facilitation, tolerance, and inhibition is designed to let students adjudicate empirically rather than by authority.

Restoration ecologists use successional theory directly when replanting degraded mine sites, coastal dunes, or wetlands: knowing that nitrogen-fixing facilitator species (legumes, alder) can jump-start soil development lets practitioners skip decades of slow natural facilitation by deliberately seeding these species first. Forestry management for timber depends on tolerance-model predictions — foresters intentionally harvest shade-intolerant pine before shade-tolerant hardwoods take over naturally, or use controlled disturbance (thinning, prescribed fire) to hold a stand at a commercially useful seral stage rather than letting it proceed to a less profitable climax composition. Invasive-species management increasingly frames the problem in inhibition terms: an invasive grass or shrub that forms a dense inhibitory sod can permanently block native tree regeneration, so restoration requires actively breaking the inhibition (mechanical removal, targeted herbicide, fire) rather than waiting for tolerance-based natural replacement that will never occur. Wetland mitigation and reclamation of strip-mined land are legally regulated in many countries using explicit successional-trajectory benchmarks — a reclaimed site must demonstrably be progressing through recognizable seral stages toward a target community within a specified number of years. In global-change biology, ecologists track how rising atmospheric CO₂, altered fire regimes, and novel invasive species are changing successional trajectories themselves — some post-fire landscapes are no longer returning toward their historical climax vegetation at all, instead locking into persistent alternative states, a live research question that directly extends the Clements–Gleason contingency debate into a warming world. Succession theory also underlies the reasoning behind the Intermediate Disturbance Hypothesis, since disturbance frequency determines how much of a landscape sits in early, mid, or late seral stage at any moment — and mid-successional communities are frequently the most species-rich, a pattern formalized when this article's ideas are extended into Species Diversity and Its Measurement.

Ask the tutor: How would the facilitation/tolerance/inhibition framework predict succession differently on a nutrient-rich secondary-succession field versus a nutrient-poor primary-succession lava flow?