The Community Concept

An ecological community is the assemblage of populations of different species that occupy a given area and interact directly or indirectly through competition, predation, mutualism, and other relationships covered under Species Interactions. This is deliberately broader than a single population's dynamics (Population Dynamics) because a community's properties — species richness, trophic structure, total biomass — emerge from many interacting populations rather than from any one of them. The contested question is not whether species co-occur — obviously they do — but whether that co-occurrence reflects a discrete, integrated unit with emergent, organism-like properties (growth, a characteristic composition, a "climax" endpoint, sharp boundaries with neighboring communities), or whether it is simply a statistical overlap of species whose individual distributions are shaped independently by physical gradients (temperature, moisture, soil, disturbance) and dispersal history. The first position is the holistic (or organismic) view; the second is the individualistic (or continuum) view. Both camps agree on the raw pattern: certain species do recur together more often than chance in similar environments — oak-hickory forest, tallgrass prairie, kelp forest. They disagree on the cause of that recurrence. Holists attribute it to tight coevolved interdependence that produces a discrete, classifiable unit with real boundaries. Individualists attribute it to shared, but not identical, environmental tolerances: species ranges overlap heavily where conditions suit many of them simultaneously, producing the appearance of a bounded community without any organizing whole enforcing it. This is exactly analogous to asking whether a rainbow's color bands are objects or an artifact of how we perceive a continuous spectrum — the underlying gradient is real, but the "bands" may be a description imposed by the observer, not a property of the gradient itself.

Step 1 — Individual species respond to individual gradients. Every species has a range of physiological tolerance for temperature, moisture, soil pH, light, and other factors, and a set of tolerances for competitors, predators, and mutualists. These tolerances differ species by species: no two species have identical requirements, dispersal ability, or response times to environmental change. This is the physiological/ecological substrate the whole debate rests on.

Step 2 — Independent responses generate independent range limits. Because tolerances differ, each species' range boundary along any given gradient (elevation, latitude, moisture) falls at a different point. If you plot the abundance of many species against a single continuous gradient, each species traces its own bell-shaped or skewed curve, peaking where conditions best match its optimum and declining toward its tolerance limits.

Step 3 — Overlap of these curves produces zones that look like "communities." Where many species' curves happen to peak or overlap in the same stretch of the gradient, an observer sees a recognizable assemblage — "oak-hickory forest" — and where curves thin out, the assemblage seems to end. But the key mechanistic point is that this apparent zonation is a statistical byproduct of many independent curves overlapping, not evidence of a coordinating unit. If the curves are checked individually (species by species), boundaries are usually staggered and diffuse rather than coincident.

Step 4 — Interactions can superimpose real structure on top of this backdrop. This is where the picture is not purely individualistic. Once species co-occur because of shared tolerances, they then interact: competitive exclusion can sharpen a boundary that would otherwise be gradual (a species is physiologically capable of persisting a bit further along the gradient but is competitively excluded there); mutualisms (mycorrhizal networks, pollination webs) and keystone predation can knit some species together more tightly than gradient overlap alone would predict; disturbance regimes (fire, storms) can reset several species simultaneously and synchronize their presence. So real community-level cohesion exists, but it rides on top of an individualistic base rather than replacing it.

Step 5 — Time adds a further wrinkle: succession. As a community assembles after disturbance, species arrive and are replaced across time as well as space. Clements viewed this successional sequence as a deterministic, repeatable developmental pathway culminating in a single stable climax community for a given climate — much like an organism's ontogeny. Gleason and later ecologists showed successional trajectories are contingent on which species happen to arrive first, local site history, and chance — closer to independent colonization histories than to a scripted developmental program.

Clements's original observations (1900s–1930s). Frederic Clements studied vegetation succession in the North American grasslands and concluded that plant communities pass through a predictable sequence of stages (pioneer, intermediate, climax) that is repeatable for a given climate, regardless of starting conditions — much as an organism develops from embryo to adult. He formalized this into the superorganism concept: a community as a discrete, bounded, coherent unit with birth (colonization), growth, maturity (climax), and even reproduction (via seed dispersal re-establishing the same climax type elsewhere). Clements's climax-formation classification became the dominant framework in American plant ecology for decades, and it strongly shaped early conservation thinking (e.g., defining "types" of natural vegetation to protect).

Gleason's rebuttal (1926, "The Individualistic Concept of the Plant Association"). Henry Gleason, working from extensive floristic survey data, argued that Clements's sharp community "types" were an artifact of coarse sampling and observer expectation. Gleason pointed out that:

• Individual plant species' ranges, when mapped carefully, rarely coincide precisely with any other species' range.
• The same set of species is almost never found in exactly the same proportions in two different stands, even within a supposedly uniform "association."
• What looks like a discrete vegetation type is simply the region where many individual species ranges happen to overlap most densely — a statistical, not organismic, phenomenon.

Gleason's paper was largely theoretical at the time and was not immediately accepted; Clements's framework dominated textbooks into the 1950s.

Robert Whittaker's gradient analysis (1950s–1960s) — the decisive quantitative test. Whittaker sampled vegetation directly along continuous environmental gradients rather than trying to classify stands into discrete types first. His most famous work mapped tree species abundance against elevation and moisture gradients in the Great Smoky Mountains (Whittaker, 1956) and in the Siskiyou Mountains of Oregon (Whittaker, 1960). What he found, quantitatively:

• Plotting the relative abundance (basal area or density) of each tree species against elevation produced individual bell-shaped curves, each with its own peak and its own upper/lower limits.
• Species' curve peaks were staggered continuously along the gradient rather than clustering at a few discrete points.
• Species' range limits (the points where a curve reaches zero) did not coincide across species — one species might drop out around 900 m while a co-occurring species persisted to 1,200 m, with no shared "edge."
• There was no statistical evidence of sharp, coincident boundaries that would mark discrete community "types"; instead, composition changed gradually and continuously — a continuum.

This method, called gradient analysis (a precursor to modern ordination techniques such as principal components analysis and detrended correspondence analysis), became the standard empirical test of the holism-versus-individualism question, and it has been repeated across many biomes — grasslands, tundra, coral reef zonation, intertidal zonation — with broadly similar results: continuous turnover is the norm, sharp community boundaries are the exception (usually associated with a genuinely sharp environmental discontinuity, like a soil-type boundary or a shoreline, rather than with community self-organization).

Fossil pollen evidence — testing the "repeatable climax" claim over deep time. If Clements were correct that a climate zone always redevelops the same discrete community, then paleoecological records should show the same species assemblages recurring together whenever similar climates recur. Pollen core studies from post-glacial North America and Europe (notably work synthesized by Margaret Davis and others from the 1960s onward) instead show that tree species migrated northward after the last glaciation at different rates and via different routes, so that the specific combinations of co-occurring tree species during migration were often novel — combinations with no modern analogue. This independent migration is strong support for the individualistic view: species track their own tolerances and dispersal opportunities rather than moving as a bound package.

Modern synthesis. Contemporary community ecology accepts the individualistic view as the correct default null model — communities are not superorganisms and rarely have sharp, universal boundaries — while recognizing genuine cohesive structure imposed by strong species interactions (keystone predation, obligate mutualisms, ecosystem engineers) on top of that gradient backdrop. The debate reframed community ecology's central question from "what type is this community?" to "how much, and by what mechanisms, does interaction structure a set of independently distributed populations?" — the question taken up directly in The Ecological Niche and Food Webs and Community Structure.

Example 1 — Reading a gradient-analysis dataset. Suppose a Whittaker-style survey along a 2,000 m elevation transect records the elevational range (lower–upper limit, meters) and abundance peak for five conifer species:

A holist expects a shared boundary — for instance, all five species dropping out near the same elevation, marking the edge of a discrete "montane forest" community and the start of a different one above it. Reading the table: the upper limits are 1,400, 1,100, 1,900, 2,000, and 2,000 m — no two species share an upper limit, and the spread is 900 m wide. The lower limits similarly range from 200 to 1,000 m. There is no elevation at which all five species simultaneously appear or disappear; instead, species drop in and out continuously across the whole 2,000 m transect. This staggered pattern is exactly Whittaker's empirical result and supports the individualistic model: what a field botanist might call "montane forest" versus "subalpine forest" is a naming convenience for a zone of maximum overlap, not a real discrete unit with a shared edge.

Example 2 — Distinguishing individualistic overlap from true interaction-driven structure. A researcher studies two competing barnacle species on a rocky shore, Chthamalus (upper zone) and Balanus (lower zone), and wants to know whether their non-overlapping distribution reflects (a) independent physiological tolerance to desiccation (an individualistic gradient effect), or (b) competitive exclusion (a true interaction effect layering structure onto the gradient). Reasoning through it: first check where each species can survive alone. If Chthamalus larvae are experimentally transplanted into the lower zone with Balanus removed, and Chthamalus survives and grows there perfectly well, then its natural absence from the lower zone in the presence of Balanus cannot be explained by desiccation tolerance alone — the physiological range extends lower than the observed range. That gap between physiologically tolerable range and observed range is the operational signature of an interaction effect. (This logic mirrors Joseph Connell's actual barnacle removal experiments on the Scottish coast, which showed Balanus physically dislodges and smothers Chthamalus in the lower zone — true competitive exclusion, not gradient sorting alone.) The lesson generalizes: a sharp boundary observed in the field is not automatic evidence against the individualistic model — it must be tested by removal/transplant experiments to see whether the boundary is set by tolerance (individualistic) or by interaction (a layer of real structure on top). This exact logical move — separating the fundamental niche from the realized niche — is developed fully in The Ecological Niche.

Henry Allan Gleason spent much of his career as a systematic botanist and curator at the New York Botanical Garden, cataloguing the flora of the northeastern United States and the Caribbean in painstaking, specimen-level detail. That taxonomic grounding — actually mapping where individual species occurred, one range at a time, rather than starting from an assumed "type" — is precisely what let him see what Clements's more top-down, classificatory approach missed: that no two species' ranges are ever quite the same, and that the "association" is a description of overlap, not an entity in its own right. Gleason published his key argument in 1926 in the Bulletin of the Torrey Botanical Club, but the paper had little immediate impact; Clements's superorganism framework, propagated through influential textbooks and applied successfully to rangeland management, remained the field's orthodoxy. Gleason revisited and sharpened his argument in a 1939 paper, "The Individualistic Concept of the Plant Association," restating the case as ecological data from careful floristic surveys accumulated in his favor. Vindication came mainly posthumously and indirectly, through Robert Whittaker's quantitative gradient analyses of the 1950s and 1960s, which supplied the rigorous numerical evidence Gleason's original argument lacked.

Conservation planning. Modern conservation increasingly avoids defining protected areas around a fixed "community type" to be preserved forever, since climate change is already causing species within any historical assemblage to shift range at different rates — an individualistic response predicted directly by Gleason's model. Instead, planners now emphasize preserving environmental gradients, connectivity corridors, and process (disturbance regimes, hydrology) so that species can independently track shifting conditions, rather than trying to hold a static species list in place.

Vegetation and habitat mapping. Ordination and gradient-analysis techniques descended directly from Whittaker's methods are the standard tools in remote sensing and GIS-based habitat classification, used to predict how vegetation zones will shift under climate scenarios by modeling individual species' response curves rather than assuming fixed community blocks move as units.

Restoration ecology. Restoration projects that try to recreate a specific historical "climax" community by planting a fixed species list often fail if the underlying environmental gradient (hydrology, soil chemistry, fire frequency) has changed; the individualistic framework redirects restoration effort toward restoring the process and gradient conditions that generate self-sustaining assemblages, letting composition emerge rather than be prescribed.

Invasive species management and novel ecosystems. Because communities are not fixed units, ecologists now routinely describe "novel ecosystems" — combinations of native and introduced species with no historical precedent — as a predictable, individualistic outcome of species tracking changed conditions independently, rather than as an aberration from some "true" community type that must be restored exactly.

Agriculture and rangeland science. Ironically, Clements's holistic, climax-based framework had major early success guiding U.S. rangeland management (predicting recovery trajectories after overgrazing), and elements of successional-stage thinking persist in silviculture and rangeland practice even though the strict superorganism model has been abandoned — a reminder that a partially wrong model can still generate useful predictions in a restricted context.

Ask the tutor: How would I design a field experiment to test whether a sharp vegetation boundary I observe on a hillside is caused by a soil discontinuity, competitive exclusion, or both?