Community development has been studied mainly in plant communities. Observations often indicate that a predictable sequence of communities follows one another:

The dominant plants change conspicuously from early to late successional stages: early successional species tend to be small, fast-growing, short-lived, opportunistic (weedy, fugitive), "r-selected" species; late successional species tend to be large, slow-growing, long-lived, equilibrium, "K-selected" species.

Primary succession: Begins in a new "lifeless" environment (e.g. lava flows, glacial retreats, sand banks, rock slides), unmodified by previous communities (e.g. no soil)

Secondary succession: Begins in an area where a previous community has been disturbed. Effects of the previous community still exist (e.g. well-developed soil) and many individuals and species from the previous community may have survived.

While primary and secondary successions often differ, typical plant seral stages include:

Common trends during progression from early to late successional stages include:

• increasing size of organisms
• increasing longevity of organisms
• shift from predominantly "r-selected" to predominantly "K-selected" species
• shift from "individualistic" to predominantly "interactive" organization
• increasing biomass
• increasing independence of physical/chemical environment
• increasing development of soils
• declining rate of change
• increasing species diversity
• increasing complexity of physical and trophic structures
• increasing habitat modification and buffering of physical/chemical extremes
• increasing complexity of energy and nutrient flows
• increasingly closed systems re-cycling of organic and inorganic materials

(N.B. the last five trends are very similar to those observed in the development of biomes)

Climax Community:

This is the final seral stage, consisting of a predictable equilibrium community that perpetuates itself indefinitely without changing into another community.

In Gleason's individualistic view, the climax is determined edaphically, by soil type and climate. In any one region, there is only one predicted climax, the characteristic biome of the region.

In Clement's interactive view, the climax is determined locally not only by soil and climate, but also by the characteristic species present and the interactions among them. This leads to a polyclimax view of a mosaic of patches of different climaxes, each of which is stable locally.

1. Facilitation model: Each seral stage modifies the physical and chemical environment in ways that make it suitable for species in the next stage to become established. These later species then eliminate the earlier successional species (now living in a less favorable environment) by competitive exclusion.

Facilitation is assumed in Gleason's individualistic concept: species become established when conditions are close to their fundamental niches, and are excluded as realized conditions no longer include their optimal fundamental niches. Repetition of this process at each stage creates the predictable, directional sequence leading towards the edaphic climax of the region.

The facilitation model appears to be a realistic description of early stages of primary succession.

2. Inhibition model: Many (even most) species characteristic of most seral stages are assumed to be already present at the start of the succession. The order in which they become dominant is determined mainly by abundance, growth rate, longevity and tolerance of unfavorable conditions. Thus, rapidly growing grasses become numerous and conspicuous very rapidly, perhaps even before the seeds of shrubs and trees have germinated, and certainly while these seedlings are sparser and smaller than the grasses. Initially one "perceives" only the grass community. Later, faster-growing shrubs become taller than the grasses, start to "inhibit" the grasses by out-competing them for light, water and nutrients, and are "perceived" as a dominant shrub community. Eventually, slowly growing large trees overtop the shrubs and become "perceived" as the new dominants.

Inhibition is also known as the "tolerance" model: species most tolerant of adverse conditions, especially intense competition, persist to become dominants of later successional stages, including the climax. The inhibition model probably explains many secondary successions that begin with numerous survivors of previous communities on developed soils.

3. Disturbance (and accumulation) model: This model assumes that recruitment of new species can occur only immediately after a disturbance, and that which species become established is largely a matter of chance determined by which species are reproductive, their powers of dispersal, distance from a source, size of the disturbance, and how quickly it fills in. Following the disturbance, a new community develops composed of survivors of the previous community, plus the new recruits.

This assemblage persists with little change until the next disturbance, when a different subset of individuals and species survive, and a different set of recruits becomes established. If small, short-lived ("r-selected") species are most likely to be killed by disturbances, while larger individuals from longer-lived ("K-selected") species are more likely to survive through repeated disturbances, members of predominantly "late successional" species will accumulate over time. An observer will see a superficial shift from predominantly early- to predominantly late-successional assemblages, but species composition will not be predictable, the process may be reversible, and there is no predictable "climax". Instead, species composition will depend largely on the frequency, nature and intensity of disturbances, and on the pools of available recruits.

The disturbance model probably acts with the facilitation model during later stages of primary succession. Both the disturbance and inhibition models probably exist and interact in most secondary successions; inhibition may dominate when disturbances are rare, but declines in importance as the frequency and intensity of disturbance increases.

This arises from Clement's interactive community concept, and the inhibition and disturbance succession models. It is visualized as a convex arc when species diversity (S; y-axis) is plotted against disturbance (x-axis). Disturbance is a gradient of increasing numbers, kinds, durations, intensities and unpredictability of disturbance. It predicts:

1. Mosaics: Compare patches of different ages that can be arrayed in a linear, temporal sequence (e.g. lava flows; fire patches)
2. Transects: Compare young to old communities across a gradient where a community is expanding to occupy new habitat (e.g. lake margins; peat bog margins)