Island Biogeography

Island biogeography is the quantitative subfield of community ecology and evolutionary biology that examines the factors governing species richness, community assembly, and endemism on oceanic islands, archipelagos, and isolated continental habitat patches. Its theoretical cornerstone is the Equilibrium Theory of Island Biogeography (ETIB), formulated by American ecologists Robert H. MacArthur and Edward O. Wilson in a 1963 Evolution paper and expanded in their seminal 1967 Princeton monograph The Theory of Island Biogeography. Moving ecology away from purely descriptive natural history toward predictive mathematical modeling, MacArthur and Wilson proposed that the number of species (S) inhabiting an island is not a static historical constant, but a dynamic equilibrium (Ŝ) maintained by the opposing processes of recurrent immigration (colonization of new species from a mainland source pool of size P) and local extinction (stochastic extirpation of established island populations).

At equilibrium, the temporal rate of change of species richness vanishes (d*S* / d*t* = I(S) − E(S) = 0), so the total number of resident species remains stable around Ŝ even as the taxonomic identity of those species continually shifts through species turnover at rate T̂ = I(Ŝ) = E(Ŝ). By linking immigration rates I(S) inversely to island isolation (distance D from the mainland) and extinction rates E(S) inversely to island area (A), the MacArthur–Wilson model provides a unified mechanistic explanation for the universal species–area relationship (S = cA^z) and forms the quantitative basis for modern conservation biology, nature reserve design, metapopulation dynamics, and habitat fragmentation studies.

Explore Global Biomes and Islands: Test your biogeographical knowledge of island endemism, oceanic archipelagos, and terrestrial ecoregions in Where in the World.

Mathematical Formulation of Immigration and Extinction Curves

Consider an island located at distance D from a continental mainland source pool containing P potential colonizing species, with current island species richness S(t) ∈ [0, P]. The rate of change of species richness with respect to time t is governed by the differential balance between the immigration rate I(S) (newly arriving species not currently present on the island, in species per unit time) and the local extinction rate E(S) (resident species going extinct per unit time):

d*S* / d*t* = I(S) − E(S)

MacArthur and Wilson derived the functional shapes of I(S) and E(S) from two first-principles ecological arguments:

1. Monotonically Decreasing Immigration Curve (∂I/∂S < 0): When an island is empty (S = 0), every arriving propagule (bird, wind-blown seed, or rafting reptile) from the mainland represents a novel species for the island, setting the maximum immigration rate I(0) = I₀. As S increases, an increasing fraction S / P of arriving propagules belong to species that are already established on the island and therefore do not add to S. Conversely, if all P mainland species reside on the island (S = P), the rate of new species addition must equal zero: I(P) = 0. In the simplest linear model where all P mainland species have equal dispersal probabilities, the immigration rate decreases linearly:

I(S) = I₀ (1 − S / P) In nature, I(S) is concave upward (∂²I/∂S² > 0) because superior dispersers (such as vagile seabirds, bats, and dandelion-plumed asteraceous seeds) colonize the island rapidly at low S, leaving only poor dispersers in the remaining uncolonized mainland pool P − S.

1. Monotonically Increasing Extinction Curve (∂E/∂S > 0): When S = 0, no species can go extinct, so E(0) = 0. As S increases, more species are present to experience stochastic environmental or demographic extinction; setting a per-species baseline extinction probability allows the simplest linear approximation:

E(S) = E_max (S / P) In reality, E(S) is concave upward (∂²E/∂S² > 0) because higher species richness S divides the island's finite carrying capacity and resource base among more species, reducing the average population size N_i per species and increasing interspecific competition and trophic cascade pressures.

Analytical Solution for Equilibrium Richness (Ŝ) and Turnover Rate (T̂)

In the linear baseline model, setting d*S* / d*t* = I(S) − E(S) = 0 yields the equilibrium species richness Ŝ and the equilibrium species turnover rate T̂:

Ŝ = I₀P / (I₀ + E_max), and T̂ = I(Ŝ) = E(Ŝ) = I₀E_max / (I₀ + E_max)

Solving the linear first-order ordinary differential equation d*S* / d*t* + [(I₀ + E_max) / P] S = I₀ from an initially barren island (S(0) = 0, such as a newly emerged volcanic island like Surtsey or Krakatau after its 1883 eruption) yields an asymptotic exponential colonization trajectory S(t) = *Ŝ*[1 − exp(−*t*/τ_eq)] toward equilibrium:

S(t) = Ŝ [1 − exp(−[(I₀ + E_max) / P] t)]

where τ_eq = P / (I₀ + E_max) is the characteristic relaxation time required for the island to reach (1 − e⁻¹) ≈ 63.2% of its equilibrium richness Ŝ.

The Distance Effect and Area Effect

By varying two geographic parameters—distance from the mainland (D, defining Near vs. Far islands) and island surface area (A, defining Large vs. Small islands)—the MacArthur–Wilson graphical model predicts four distinct equilibrium combinations:

• The Distance (Isolation) Effect: Passive or active dispersal across an inhospitable ocean or matrix declines with distance D (following an exponential decay I₀(D) ∝ exp(−α*D*)). Consequently, a Near island has a higher immigration curve I_near(S) > I_far(S) than an identical Far island, resulting in higher equilibrium species richness (Ŝ_near > Ŝ_far) and higher turnover (T̂_near > T̂_far).
• The Area Effect: A Large island supports larger total population sizes (N_i ∝ A), buffering resident populations against demographic stochasticity, inbreeding depression, and localized disturbances (such as hurricanes or droughts), while also providing greater topographic and habitat heterogeneity. Thus, a Large island has a lower extinction curve E_large(S) < E_small(S) than a Small island, resulting in higher equilibrium species richness (Ŝ_large > Ŝ_small) but a lower per-species turnover rate (T̂_large < T̂_small).

Arrhenius Power Law and Canonical z-Exponents

Long before 1967, field naturalists including H. C. Watson (1859) and Swedish physical chemist Svante Arrhenius (1921) observed that the number of species S recorded in a survey increases systematically with sampled area A according to a power-law scaling equation:

S = c A^z ⇔ log₁₀ S = log₁₀ c + z log₁₀ A

where c is a taxon- and biogeographic-region-specific constant representing species density per unit area (A = 1), and z is the dimensionless scaling exponent (the slope on a log–log plot). Using a lognormal distribution of species abundances (Frank Preston's 1962 canonical lognormal hypothesis) coupled with island colonization–extinction equilibrium, MacArthur and Wilson demonstrated theoretically that isolated islands should converge on a canonical exponent of z ≈ 0.263.

Empirically, a vast meta-analysis of plant, bird, reptile, and insect datasets across global archipelagos reveals three distinct scaling regimes for the slope z:

1. Non-Isolated Continental Sample Plots (z ≈ 0.12–0.18): Within a continuous mainland forest or grassland, small plots still receive a constant influx of transient individuals from immediately adjacent habitat (the rescue effect and spatial mass effect), keeping local extinction low and flattening the species–area slope.
2. True Oceanic Islands and Isolated Habitat Fragments (z ≈ 0.20–0.35): Across island archipelagos (such as Galápagos land plants, West Indian herpetofauna, or East Indies birds) and isolated "sky island" mountaintops, the absence of a continuous rescue effect steepens the slope to z ≈ 0.25–0.35. As a famous rule of thumb formulated by P. J. Darlington (1957) for z ≈ 0.301 (since 10^0.301 = 2), a tenfold (90%) reduction in island or habitat area results in an approximately 50% reduction in equilibrium species richness.
3. Inter-Provincial / Inter-Continental Scales (z ≈ 0.60–1.0): Comparing entire biogeographic realms with separate evolutionary histories of speciation yields the steepest slopes.

The Simberloff–Wilson Mangrove Defaunation Experiment (1969)

To test the Equilibrium Theory of Island Biogeography directly in the field, Daniel S. Simberloff and Edward O. Wilson (1969–1970) conducted a classic manipulative experiment on six small red mangrove (Rhizophora mangle) islets (11–18 m in diameter) in Florida Bay, located at varying distances (2 m to 533 m) from larger source keys. After conducting a complete baseline census of all resident arboreal arthropod species on each islet (typically 20–45 species of insects and spiders), Simberloff and Wilson enclosed each entire island under a plastic tent and defaunated the arthropods using short-lived methyl bromide fumigation without harming the mangrove trees.

Monitoring every arriving and disappearing arthropod species at regular intervals over two years confirmed all core predictions of the MacArthur–Wilson model:

• Species richness S(t) on all defaunated islands climbed rapidly and plateaued within 150–250 days back at their pre-defaunation equilibrium values Ŝ (following a transient slight overshoot prior to predator arrival).
• The nearest island (E2, 2 m from a large source key) recovered the highest equilibrium richness (Ŝ ≈ 43 species), whereas the most distant island (E1, 533 m away) stabilized at the lowest richness (Ŝ ≈ 25 species).
• Even after S(t) stabilized at Ŝ, taxonomic composition continued to turn over dynamically at a rate of T̂ ≈ 0.5–1.5 species per day, confirming that constancy of species number reflects a dynamic balance of immigration and extinction rather than a static community.

The General Dynamic Model (GDM) and Taxon Cycles

On very large, remote oceanic islands (such as the Hawaiian, Canary, or Galápagos archipelagos) where the immigration rate I₀ from distant continents is extremely low, biological communities do not reach equilibrium solely via immigration; instead, in situ cladogenetic speciation and adaptive radiation (G(S)) become major sources of new species, modifying the balance equation to d*S* / d*t* = I(S) + G(S) − E(S). In 2008, Robert J. Whittaker and colleagues integrated plate tectonics hotspot island ontogeny with MacArthur–Wilson dynamics in the General Dynamic Model (GDM) of Oceanic Island Biogeography ( the ATT² model, linking diversity to island Area, Time, and Time²). As a volcanic hotspot island emerges from the ocean, grows in area and topographic elevation to middle-aged maturity, and then subsides and erodes over 5–10 million years into a low-lying atoll and seamount, carrying capacity and habitat heterogeneity follow a humped parabolic trajectory over geological time, driving peak speciation rates and endemic diversity on middle-aged islands (such as Maui and Oahu) before extinction dominates on old, eroding islands (such as Kauai and Midway).

Habitat Fragmentation, Extinction Debt, and Reserve Design

In the 1970s, Jared Diamond (1975), John Terborgh, and Michael Soulé recognized that human deforestation, agriculture, and urbanization transform continuous continental ecosystems into isolated "habitat islands" surrounded by an inhospitable anthropogenic matrix. Applying MacArthur–Wilson principles to conservation reserve design established several foundational rules:

• Area and Edge Effects: Larger nature reserves support larger population sizes and lower extinction rates E(S) than small fragments, while minimizing the perimeter-to-area ratio (edge habitats exposed to microclimatic desiccation, invasive species, and nest predation).
• Connectivity and Wildlife Corridors: Stepping-stone patches and vegetated wildlife corridors effectively reduce the inter-patch isolation distance D, increasing inter-patch immigration I(S) and creating a rescue effect (Brown & Kodric-Brown, 1977) where gene flow and immigrant individuals prevent local populations from going extinct.
• Relaxation and Extinction Debt (Tilman et al., 1994): When a large continuous forest of area A₀ and richness S₀ = cA₀^z is rapidly cleared down to a small remnant fragment of area A₁ ≪ A₀, the surviving species richness inside the remnant does not drop instantaneously to the new equilibrium Ŝ₁ = cA₁^z. Instead, during the relaxation period (I < E), the fragment temporarily holds more species than its reduced area can sustain long-term; the difference S₀ − Ŝ₁ represents an unpaid extinction debt that will go extinct over subsequent decades unless habitat area or connectivity is restored.

Related concepts include Adaptive Radiation and Keystone Species.

What was the SLOSS debate in conservation biology and how was it resolved?

Triggered by Jared Diamond's (1975) application of island biogeography to nature reserves and challenged by Daniel Simberloff and Lawrence Abele (1976), the SLOSS debate asked whether a Single Large Or Several Small reserves of identical total cumulative area preserve more species. Empirical and theoretical synthesis showed there is no universal one-size-fits-all answer: a Single Large reserve is essential for preserving area-sensitive apex predators, large-bodied vertebrates with large home ranges, and interior forest specialists vulnerable to edge effects, and it minimizes stochastic local extinction E. Conversely, Several Small reserves distributed across different environmental gradients can capture higher beta diversity (spatial species turnover between habitats) and guard against a single localized catastrophe (such as a wildfire or disease outbreak) wiping out an entire species, particularly when the small reserves are linked by dispersal corridors.

What is the "Island Rule" (Foster's Rule) in evolutionary biology?

Formulated by J. Bristol Foster (1964) and expanded by Leigh Van Valen (1973), the Island Rule describes a recurring ecogeographical trend in natural selection on islands: following colonization of an isolated island with depauperate predator guilds and low interspecific competition, small-bodied mainland lineages evolve toward gigantism (such as the dodo bird of Mauritius, Galápagos giant tortoises, and giant Canary Island rats), whereas large-bodied mainland mammals evolve toward insular dwarfism due to severe resource limitation during seasonal famines (such as the 1.8-m-tall Pleistocene dwarf mammoths and pygmy elephants of Mediterranean and California Channel Islands, and the hominin Homo floresiensis).

1. MacArthur, R. H., & Wilson, E. O. (1963). "An Equilibrium Theory of Insular Zoogeography." Evolution, 17(4), 373–387. doi:10.1111/j.1558-5646.1963.tb03295.x
2. MacArthur, R. H., & Wilson, E. O. (1967). The Theory of Island Biogeography. Monographs in Population Biology, Vol. 1. Princeton University Press. https://press.princeton.edu/books/paperback/9780691088365/the-theory-of-island-biogeography
3. Simberloff, D. S., & Wilson, E. O. (1969). "Experimental Zoogeography of Islands: The Colonization of Empty Islands." Ecology, 50(2), 278–296. doi:10.2307/1934856
4. Whittaker, R. J., Triantis, K. A., & Ladle, R. J. (2008). "A general dynamic theory of oceanic island biogeography." Journal of Biogeography, 35(6), 977–994. doi:10.1111/j.1365-2699.2008.01892.x