Mid-Domain Effect

The mid-domain effect (MDE) is a quantitative biogeographical hypothesis and geometric null model proposing that when the geographic ranges of species are distributed randomly within a bounded spatial domain—free of any environmental, climatic, or productivity gradients—the hard boundaries of the domain constrain the placement of large ranges and cause species ranges to overlap most densely near the geometric center, generating a unimodal, parabolic peak in biodiversity (Colwell & Hurtt, 1994; Colwell & Lees, 2000). Formulated by American ecologists Robert K. Colwell and George C. Hurtt in 1994 and named the "mid-domain effect" by Robert K. Colwell and David C. Lees in 2000, the theory provides a geometric baseline for evaluating whether empirical patterns of species richness across latitude, elevation, and ocean depth reflect biological sorting along environmental gradients or the spatial geometry of continents, mountain slopes, and ocean basins.

Within macroecology and the study of the biosphere, species distributions are bounded by physical barriers where habitable terrain terminates: continents end at coastlines, terrestrial elevation transects are bounded by sea level at their base and mountain summits at their peak, and marine bathymetric gradients are bounded by the air–sea interface and the abyssal seafloor. Just as tossing sticks of assorted lengths randomly into a narrow box results in more sticks overlapping the middle of the box than either end, geometric boundary constraints enforce a non-uniform expectation of species richness across any bounded biogeographical domain.

Continuous One-Dimensional Analytical Models

To isolate pure geometry from abiotic sorting, consider a one-dimensional spatial domain normalized to the unit interval x ∈ [0, 1], representing a latitudinal transect between two continental coasts or an elevational transect from sea level (x = 0) to a mountain summit (x = 1). Suppose each species i (i = 1, 2, ..., S) has a continuous, unfragmented geographic range of span r_i ∈ (0, 1) and range midpoint m_i.

For a species range to remain inside the habitable domain [0, 1] without crossing the lower boundary (x = 0) or upper boundary (x = 1), its lower limit (m_i − r_i / 2) must be at least 0 and its upper limit (m_i + r_i / 2) cannot exceed 1. Consequently, for any given range size r_i, the permissible interval of range midpoints is constrained to:

• Permissible Midpoint Interval: m_i ∈ [ r_i / 2, 1 − r_i / 2 ]
• Width of Permissible Midpoint Zone: L(r_i) = 1 − r_i

In Colwell and Hurtt's (1994) Model 2 (further analyzed analytically by Michael R. Willig and S. Kathleen Lyons in 1998), if both the lower and upper range limits are drawn from uniform random distributions on [0, 1], the probability P(x) that a randomly generated species range covers a focal point x is the product of the probability that one range endpoint falls below x and the other falls above x:

• Single-Species Overlap Probability: P(x) = 2x(1 − x)
• Expected Species Richness at Position x: E[S(x)] = 2 S_total x(1 − x)
• Binomial Variance of Richness at Position x: Var[S(x)] = S_total [ 2x(1 − x) ] [ 1 − 2x(1 − x) ]

This analytical equation describes a symmetric parabola that equals zero at the boundaries (x = 0 and x = 1) and reaches its maximum richness at the geometric midpoint (x = 0.5), where expected richness equals half of the regional species pool (E[S(0.5)] = 0.5 S_total).

Empirical Range-Size Frequency Distribution (RSFD) Randomizations

Because natural biotas rarely possess uniformly distributed range sizes—most taxonomic clades follow a right-skewed hollow curve with many narrow-ranged endemics and few wide-ranged generalists, as studied under Rapoport's rule—Colwell and Lees (2000) developed Monte Carlo randomization algorithms (implemented in the software package RangeModel) that retain the observed range-size frequency distribution (RSFD) of an assemblage while randomizing range placement:

1. Conditional Probability per Range Size (r_i): When a species with fixed empirical range span r_i has its midpoint m_i drawn uniformly from [r_i / 2, 1 − r_i / 2], the probability P(x | r_i) that its range overlaps point x is a piecewise trapezoidal function of x with denominator (1 − r_i).
2. Role of Large-Ranged vs. Small-Ranged Taxa: Narrow-ranged species (r_i ≪ 1) can be placed almost anywhere along [0, 1] and generate a flat plateau across most of the domain. Conversely, wide-ranged species (r_i exceeding 0.5) are forced to overlap x = 0.5 on every randomization trial, driving the curvature of the mid-domain peak.

1. Elevational Diversity Gradients (Mid-Elevation Humps)

Synthesis of complete mountain transects by Carsten Rahbek (1995, 2005) demonstrated that a mid-elevation peak ("hump-shaped" pattern) between 800 m and 2,000 m is the most common elevational diversity pattern worldwide (~50% of studies), outnumbering monotonic declines from lowland tropical rainforests to alpine peaks. The mid-domain effect provides a direct geometric baseline for why diversity peaks at intermediate altitudes: species adapted to mid-elevations can disperse both upslope and downslope, whereas lowland ranges are truncated at sea level and alpine ranges are truncated at the mountain summit. In studies of epiphytic ferns, bryophytes, small mammals, and ants along Neotropical and Paleotropical mountains (Colwell et al., 2004; McCain, 2007), empirical richness regressions against 1D MDE predictions yield R² values between 0.50 and 0.85, particularly when combined with mid-elevation cloud-zone humidity peaks.

2. Latitudinal and Continental Gradients (Madagascar and Africa)

While the global latitudinal diversity gradient across both hemispheres requires climatic and evolutionary explanations beyond pure geometry, coastline boundaries strongly constrain range overlap within bounded landmasses and islands:

• Madagascar Rain Forest Biome (Lees, Kremen, & Andriamampianina, 1999): Along the 1,350 km north–south rain forest belt of eastern Madagascar, 2D geometric null models explained 65% to 92% of the mid-island richness peak across 913 species of butterflies, frogs, birds, and lemurs.
• Sub-Saharan African Birds (Jetz & Rahbek, 2001): Using a two-dimensional spreading-dye null model across continental Africa, Walter Jetz and Carsten Rahbek showed that the mid-domain effect strongly predicts the richness pattern of wide-ranging bird species, whereas narrow-ranged endemic species cluster along topographically complex East African montane rift valleys where speciation and microclimatic stability peak.

3. Marine Bathymetric (Depth) Gradients

In marine ecosystems, benthic invertebrates (polychaetes, gastropods, and bivalves) and demersal fishes consistently exhibit a unimodal peak in species diversity at bathyal mid-depths (~1,500 m to 2,500 m) along the continental slope, declining both upward toward the continental shelf and downward toward the 4,000–5,000 m abyssal plain (Pineda & Caswell, 1998). Because the sea surface and abyssal seafloor act as vertical boundaries for depth-zoned larvae, geometric MDE constraints combine with organic carbon export flux from the biological pump to shape deep-sea diversity profiles.

Critics of the mid-domain effect (Zapata, Gaston, & Chown, 2003) noted that in nature, most species ranges terminate before hitting coastlines or summits because physiological limits—governed by Shelford's Law of Tolerance—or biotic interactions (apparent competition and interspecific competition) prevent further range expansion, and that empirical range-size distributions already reflect evolutionary history.

In response, Colwell, Rahbek, and Gotelli (2004) demonstrated that geometric constraints and environmental gradients are complementary: the MDE represents the spatial null expectation that can be incorporated as an offset covariate inside spatial generalized linear models alongside temperature, water availability, and ecological assembly rules within a regional metacommunity.

1. Colwell, R. K., & Hurtt, G. C. (1994). "Nonbiological gradients in species richness and a spurious Rapoport effect." The American Naturalist, 144(4), 570–595. https://doi.org/10.1086/285695
2. Colwell, R. K., & Lees, D. C. (2000). "The mid-domain effect: geometric constraints on the geography of species richness." Trends in Ecology & Evolution, 15(2), 70–76. https://doi.org/10.1016/S0169-5347(99)01767-X01767-X)
3. Jetz, W., & Rahbek, C. (2001). "Geometric constraints explain much of the species richness pattern in African birds." Proceedings of the National Academy of Sciences, 98(10), 5661–5666. https://doi.org/10.1073/pnas.091100998
4. Colwell, R. K., Rahbek, C., & Gotelli, N. J. (2004). "The mid-domain effect and species richness patterns: what have we learned so far?" The American Naturalist, 163(3), E1–E23. https://doi.org/10.1086/382056