The forest architecture hypothesis for diversity maintenance.

The forest architecture hypothesis for diversity maintenance C an tree species coexist by differing in size? T. Kohyamai recently put forward a convincing case that they can. While there is no shortage of plausible hypotheses for coexistence of competing species, few have received satisfactory tests. The 'forest architecture' hypothesis of Kohyama, however, is refreshing in that it has received a strong test soon after formulation. The nature of the hypothesis and the approach to testing it using models and field data merit close examination. Kohyama formulates the forest architecture hypothesis in terms of a mathematical model designed to explain coexistence of three numerically dominant, non-pioneer tree species in the warm temperate rainforest of Yakushima Island, southern Japan. Examination of this model reveals four key elements: (a) one-sided competition for light, with larger individuals affecting smaller individuals (but not vice versa), independent of species identities; @) segregation of species by height so that the species tend to occupy different layers of the forest when mature (on Yakushima Island the layers are canopy , subcanopy and understorey); (c) a trade-off across species between mature height and survival, reproduction or re-generative ability; and (d) spatio-temporal variation in size structure due to the opening and closing of gaps. To understand how these elements lead to coexistence, note from (a) that mature canopy trees will experience intra-specific competition for light because of height differences within the species, even though they will not experience interspecific competition because of (b). Such intraspecific competition is likely to place an upper limit on the density of canopy trees. The canopy will not capture all the light, but the light that does pass through may not be sufficient to support a population of trees of shorter stature. A canopy species with a high mortality rate, or low reproductive or re-generation rate, is likely to be limited at a level that lets substantial light through to lower layers, as such a species needs high light per unit leaf area to make up for its unfavourable demographic parameters. The trade-off (c) means that species of shorter stature can make do with lower light per unit leaf area, and therefore have the possibility of maintaining a population below the canopy. A subcanopy species will be limited both by interspecific competition from the layer above, and by intraspecific competition. It may also be limited at a level that allows sufficient light to penetrate below …