When animals and plants colonise an island, the biodiversity of endemics relies on distance from a mainland and the area of the island itself. This means that remote islands maintain higher biodiversity in general. Even outstanding endemics such as the little dodo (the national bird, the manumea) in Samoa need to fit into a pattern.

What is of great interest is how groups of islands seem to combine their areas (Surface Area Relationships or SARs), especially with small islands in archipelagos. This is naturally called the small island effect, with a good example in the Kapingamarangi Atoll in Micronesia, to the north of Papua New Guinea where plants are significantly different from their relatives. The authors of this paper also studied biodiversity in the West Indian archipelagos and the British Isles with the results displayed within.

Because of their instability in terms of environmental conditions, extinction rates don't relate to area on small islands. Species diversity rarely stabilises where it has been thought that great storms can wipe out the entire community. The authors of a new paper today suggest that immigration is the key factor so far neglected by theorists. Ryan A. Chisholm, Tak Fung, Deepthi Chimalakonda and James P. O'Dwyer supply a thesis on the Maintenance of biodiversity on islands in the Proc. Roy Soc B.

Large islands, instead of recovering more quickly from environmental disturbance, often seem to have lower diversity despite a likely higher immigration rate. Instead of the theories noted above, niche constraints and immigration are proposed as maintaining the very high diversity sometimes achieved on small (and on large) islands.

This means that with niche diversity increasing slowly with area, niche constraints dominate on small islands while immigration dominates on big islands. Such a unified theory on island diversity needs severe testing, with mathematical models now available in both complex and simplified form. Many animal and plant groups were tested alongside a broad range of archipelago types. Key to the theory is the classic prediction. In this case the theory predicts that high immigration rate would decrease in a critical area where there is a transition between niche-structured regimes on small islands and colonization-extinction regimes on larger islands.

As an example of this prediction, birds and plants migrate to islands more easily than mammals in general. Mammals have a critical area for this transition at 20km² with birds at 0.78km²! Also, total species richness is closely associated with niche diversity. A small island in Oceania might have a tree niche, a grassland niche, and a salty shrub niche, for example. These 3 niches would give 3 species a long term future on the island, although in reality these numbers would be much higher hopefully. Low immigration would maintain this relationship, though species would change over time.

Even general ecologies can be fitted into these new theories. With much greater immigration, critical areas would be very small with hundreds of species coexisting in a rainforest. This requires treating a mainland as simply a very big island, which seems realistic. In aquatic systems, lakes and even isolated coral reefs fit neatly into the theory, in many cases with proximity to a continental shelf a critical factor.

As far as conservation is concerned, forest fragmentation is creating tiny pieces of habitat that will function as islands. Biodiversity would naturally be lost from such entities because immigration is reduced and, secondarily, because of niche structural change, according to the theory. Species loss would be huge even if the niches remain more or less intact. The niche here is a dominant force while immigration is more dominant where larger islands are conserved.