Module 1: The Seven Biogeographic Realms

The seven canonical terrestrial realms, together with the circumpolar Antarctic, are summarised below. Areas are rounded to the nearest 0.1 million km\(^2\)and species counts to nearest hundred for vertebrates.

In the Holt et al. (2013) 11-realm scheme, Madagascar, the Panamanian transition zone, and the Saharo-Arabian are split from the Afrotropical and adjacent regions. The seven Sclater–Wallace realms remain recognisable as “super-realms”.

No boundary in terrestrial biogeography has been as sharply drawn, or as repeatedly redrawn, as the one separating Asian and Australian faunas in the Malay Archipelago. Three principal lines are recognised:

• Wallace’s Line (1859): runs between Bali and Lombok and between Borneo and Sulawesi. Marks the eastern limit of the Sunda Shelf and of typically Asian placental mammals.
• Weber’s Line (1902): the line of “faunal balance”, east of Sulawesi, where Asian and Australian elements contribute roughly equally.
• Lydekker’s Line (1896): marks the western edge of the Sahul Shelf (Australia + New Guinea). East of Lydekker’s Line the fauna is essentially fully Australian.

Between Wallace’s and Lydekker’s Lines lies Wallacea: Sulawesi, the Lesser Sundas, Moluccas, and Timor. Wallacea is a transitional filter zone in which neither continental biota dominates, because the intervening straits have remained deep even at glacial low-stands of sea level.

The modern phylogenetic analysis by Holt et al. (2013) recovers the three lines as discontinuities in pairwise \(\beta_{\text{phy}}\)-diversity. The sharpest break is Lydekker’s, reflecting the long (~80 Myr) isolation of the Sahul block from Sundaland.

Ben Holt and colleagues, in a 2013 Science paper (339:74–78), performed the first global, species-level, phylogenetic reconstruction of biogeographic realms. The dataset comprised 21,037 species of amphibians, birds, and non-marine mammals with time-calibrated phylogenies and spatial range maps. Their metric was phylogenetic \(\beta\)-diversity:

Using UPGMA clustering on the \(\beta_{\text{phy}}\) matrix, Holt et al. recovered 11 realms with 20 subrealms, statistically supported at the deeper nodes. The main departures from Sclater’s six are:

• Madagascan realm separated from Afrotropical (phylogenetic \(\beta \approx 0.9\) across the Mozambique Channel).
• Panamanian transition zone separated from both Neotropical and Nearctic.
• Saharo-Arabian realm separated from Afrotropical (shares more with Palearctic at deep nodes).
• Sino-Japanese subrealm separated from wider Palearctic.
• Oceanian recovered as its own super-realm, reflecting long-distance overwater colonisation by a few lineages.

The central theoretical claim of Holt et al.’s paper is that realm boundaries are information-dense: they carry signatures of plate tectonics, sea-level history, and climatic barriers reaching back tens of millions of years. Because phylogenetic metrics weight each species by its evolutionary distinctiveness, rather than counting species uniformly, deep splits are amplified and shallow human-disturbance changes are down-weighted.

Two mutually-exclusive explanations are available for any disjunct distribution:

• Vicariance: an ancestral continuous range was split by the appearance of a barrier (mountain uplift, sea-level rise, continental rifting). The two descendant ranges date to the barrier.
• Dispersal: the range began in one region and subsequently crossed the barrier. The daughter lineage dates to the crossing, not the barrier.

The two hypotheses make distinct predictions about the age structure of the descendants. Under vicariance, the divergence date equals the barrier age. Under dispersal, it is younger. With dated molecular phylogenies, one can in principle distinguish them:

Classic examples in which both processes have left their fingerprints:

• Ratite birds (ostrich, emu, rhea, kiwi, moa, cassowary, tinamous): once thought purely Gondwanan vicariant, now dated as mostly post-Gondwanan flighted dispersers with secondary flightlessness (Mitchell et al. 2014).
• Madagascan vertebrate fauna: Madagascar rifted from Africa at ~160 Ma but most endemic lineages (lemurs, tenrecs, euplerid carnivorans) arrived 30–60 Ma by oceanic rafting, i.e. dispersal not vicariance (Ali & Huber 2010).
• Southern beeches (Nothofagus): present in South America, Australia, New Zealand, New Guinea. Divergence dates are broadly consistent with Gondwanan vicariance plus some transoceanic dispersal.

Endemic species are those restricted to a single biogeographic region. The pattern of endemism across realms is shaped by (i) area, (ii) isolation time, (iii) topographic and climatic heterogeneity, and (iv) extinction history. The multiplicative relationship is summarised as:

The across-realm Arrhenius exponent \(z\) for endemic vertebrates is about 0.2 (Kier et al. 2009). Two realms deviate strongly from the area scaling: Madagascar (vastly more endemic than its area predicts, due to 88 Myr isolation) and Antarctica (vastly fewer, due to Miocene glaciation). The simulation in this module recovers these residuals numerically.

Some realms preserve lineages long extinct elsewhere. Three classic cases:

• Madagascan lemurs: five extant families (Cheirogaleidae, Indriidae, Lemuridae, Lepilemuridae, Daubentoniidae) all descended from a single Eocene colonisation event ca 50–65 Ma (Yoder 2013). Lemurs never lived in continental Africa in the modern sense, making Madagascar the entire distribution of the clade.
• New Zealand tuatara (Sphenodon punctatus): the sole surviving member of Rhynchocephalia, sister order to Squamata, which otherwise vanished from the fossil record at the end of the Cretaceous. The tuatara lineage diverged from squamates ca 240 Ma.
• Australasian monotremes (Ornithorhynchusand echidnas): egg-laying mammals, sister to therians (marsupials + placentals), and the only surviving lineage of an otherwise Mesozoic radiation.

Relictual faunas are heavily concentrated in realms with long isolation and relatively low taxonomic turnover. They constitute disproportionate contributions to global phylogenetic diversity (PD): a single tuatara species contributes more PD than many entire squamate genera combined, because its branch is so long and so isolated.

Within each terrestrial realm, species richness peaks in the tropics and declines towards the poles—the latitudinal diversity gradient (LDG). The gradient is present in both realm-specific subsets and the global pool, and is visible for virtually every well-sampled clade from flowering plants to beetles to amphibians.

Among the competing explanations for the LDG, three have the strongest empirical support and continue to be tested at the realm scale:

• Species-energy hypothesis: species richness scales with available energy flux (Wright 1983; Hawkins et al. 2003). Tropical realms see higher incoming shortwave flux and longer growing seasons.
• Time-integrated hypothesis: tropical realms have been climatically stable over longer intervals than temperate realms, accumulating more lineages (Fine 2015).
• Area hypothesis: integrating across realms, tropical biomes occupy larger, more contiguous areas (Rosenzweig 1995).

The interaction with realm structure is important. The Palearctic and Nearctic extend further into the temperate zone than the tropical realms; if one compared equal-area bands, the LDG would partly collapse—tropical realms are richer because they are large and tropical, not only because they are tropical.

Realm boundaries are not fixed. Plate tectonics reconfigures connections on tens-of-Myr timescales; sea-level cycles open and close intermittent corridors; mountain uplift closes rain-shadow barriers. The major post-Cretaceous events include:

• Great American Biotic Interchange (GABI, starting ~3 Ma): closure of the Panama Isthmus exchanged faunas between North and South America. Xenarthrans (sloths, glyptodonts, armadillos) moved north; procyonids, canids, and eventually felids and equids moved south. Asymmetric survival: most North-to-South migrants persisted, most South-to-North did not.
• Miocene “gomphothere exchange” (~20 Ma): initial faunal exchange between Afro-Arabia and Eurasia via the Gomphotherium Landbridge (closure of Tethys), bringing African elephants, apes, and rodents into Eurasia.
• Beringian dispersal pulses: Pleistocene glacial low-stands exposed a Bering land bridge, allowing mammalian exchange between Nearctic and Palearctic (mammoths both ways, humans east, bison west, horses extinct from the Americas after the Last Glacial).

These events inject noise into any “equilibrium” reading of realms: a large fraction of North American mammal genera derive from the Old World, and vice versa. The Holt et al. phylogenetic analysis partially disentangles this by weighting deep nodes more heavily than recent dispersers.

\(\beta\)-diversity is the variation in species composition between sites. For two sites with \(|A|\) and \(|B|\) species and \(|A\cap B|\) in common, three widely-used metrics are:

Koleff et al. (2003) catalogued 24 \(\beta\)-diversity metrics, which partition into turnover components and nestedness components (Baselga 2010). Simpson’s \(\beta_{\text{sim}}\) is insensitive to nestedness and is preferred when one wants to isolate pure turnover—a suitable choice for comparing deeply separated realms where richness contrasts are not the question of interest.

For phylogenetic realms (Holt et al. 2013), the analogue of \(\beta_{\text{sim}}\)weights branches rather than tips:

Madagascar is the paradigmatic case for promoting a subregion to full realm status. The island separated from Africa 160–170 Ma (Jurassic rifting) and from India 88 Ma (late Cretaceous). At 587,000 km\(^2\) it is smaller than Texas, yet it holds about 900 endemic vertebrate species, five endemic primate families, two endemic bird families (mesites, ground-rollers), and about 90% of reptile species found nowhere else on Earth.

Holt et al.’s phylogenetic \(\beta\) between Madagascar and the nearest African mainland exceeds 0.9 for amphibians and mammals; this is comparable to the Palearctic–Australasian contrast despite the two being separated by only 400 km of ocean. The explanation is the combination of a deep phylogenetic filter (only rafters and long-distance flyers crossed) and an unusually long in situ radiation window (88 Myr).

Yoder et al. (2003) and Ali & Huber (2010) proposed that Paleogene ocean currents around Madagascar were favourable for rafting from East Africa (eastward Mozambique Channel currents before the Eocene), accounting for the observed pulse of arrivals in the 30–65 Ma window. Since the Miocene the current reversed, reducing colonisation and allowing endemism to accumulate without significant dilution.

Myers et al. (2000, Nature 403:853) defined biodiversity hotspots as regions with ≥ 1,500 endemic vascular plant species (0.5% of world total) that have lost ≥ 70% of their primary vegetation. The original 25 hotspots covered 1.4% of land area yet held 44% of plant and 35% of vertebrate endemics.

Hotspots map strongly onto realm boundaries: the majority are in the Neotropical (Atlantic Forest, Tropical Andes, Mesoamerica), Afrotropical (Cape Floristic, Guinean Forests), Indomalayan (Sundaland, Western Ghats), and Madagascar realms. This is not coincidental; realms with long isolation concentrate endemism, and isolated regions are often also disproportionately disturbed by humans.

The conservation implication is that prioritisation at the global scale should weight phylogenetic distinctiveness as well as raw species counts. A lost tuatara represents far more phylogenetic diversity than a lost species of a recently-radiated clade; EDGE (Evolutionarily Distinct and Globally Endangered) metrics (Isaac et al. 2007) formalise this.

Antarctica is biogeographically anomalous: ~14 million km\(^2\) of terrestrial area, of which 98% is ice-covered, supports only two flowering plant species (Deschampsia antarctica and Colobanthus quitensis), about 100 mosses, 350 lichens, and no terrestrial vertebrates. Marine and air-breathing vertebrates (petrels, skuas, penguins, pinnipeds, cetaceans) colonise the continent seasonally or permanently but are not strictly terrestrial.

The realm acquired its present biota largely after the Miocene cooling (ca 15 Ma) that converted a Nothofagus-Araucaria forested continent into a polar desert. Pre-glacial Antarctic fossils document an austral temperate flora shared with southern South America and Australasia—the deep Gondwanan signature invoked in Hooker’s 1859 essay. The modern Antarctic terrestrial biota is therefore a post-glacial relict rather than a continuous representation of the original Gondwanan inheritance.

In Holt et al.’s phylogenetic analysis, Antarctica is often excluded from \(\beta\)-diversity calculations because of its depauperate vertebrate fauna; the continent is effectively a separate realm by historical geography but with so few taxa that statistical methods become degenerate. In the present module’s endemism simulation, Antarctica appears as a strongly negative residual—it holds far less endemism than its area would predict.