POWER FRACTION - A NEW EXPLANATION OF RELATIVE ABUNDANCE PATTERNS IN SPECIES-RICH ASSEMBLAGES

Analysis of species abundance patterns in assemblages with relatively large number of species has been an important issue in community ecolo gy for several decades. Following Preston's pioneering work, Sugihara proposed a model to account for such patterns in a diverse range of co mmunities, which has been given further support from analyses of recen t data that are considered to be of particularly high quality. This pa per re-examines Sugihara's approach and points out that there has been confusion and misunderstanding among workers as to the exact nature o f his widely-publicised hypothesis. In particular, the 'fixed ratio' d ivision cannot be considered as an expected (average) pattern of the t riangular assumption; they represent fundamentally different entities. Further, Tokeshi's Random fraction model should not be treated as syn onymous with Sugihara's fixed division model. The RF model has its own identity as a niche apportionment model and can account for patterns in some species-rich assemblages. With this background, this paper pro poses a new niche apportionment model to explain relative abundance pa tterns in ecological communities, termed the Power fraction model. The PF model envisages that the probability (p) of selection for a subseq uent division is positively but weakly related to niche sizes/abundanc es (x) of species as a power function (p alpha x(k) where 0 less than or equal to k less than or equal to 1.0; i,e. niche division/invasion is more likely in species with high abundance/large niche) and that di vision occurs with any ratio (i.e. a barrier to split a species' popul ation may occur anywhere in the species' range), with the RF model rep resenting an extreme case on the spectrum of the PF model. The Power f raction model with k similar to 0.05 or k similar to 0-0.2 demonstrate s a good fit to a miscellany of data from species-rich assemblages. Th e PF model is not restrictive in terms of division ratios, and can be framed as either a community-specific or a global explanation of patte rns. Thus, the new model is more flexible and realistic from ecologica l and evolutionary points of view and offers a possibility of cross-co mmunity comparisons within a uniform, integrative framework.