DISTINGUISHING AMONG THE FOUR OPEN HYPOTHESES FOR LONG-TERM TRENDS IN ECOSPACE DIVERSIFICATION: A NULL MODEL APPROACH

Paleoecologists have long theorized three dominant trends in ecospace occupation during the Phanerozoic, each resulting ultimately from an increase in assemblage-wide (alpha) species richness. Redundancy posits that ecospace is largely unchanged, with successive species occupying life habits similar to those previously occupied. Partitioning posits that ecospace is subdivided as species progressively occupy intermediate life habits. Expansion posits that ecospace is increased as species progressively occupy more extreme habits. These three hypotheses should result in driven trends, each following distinct assembly rules. A fourth hypothesis, in which ecospace occupation results from the independent accumulation of species, has not been sufficiently considered and serves as a passive null model.

To model these hypotheses, life habits are operationalized using a 44-dimensional ecospace framework universally applicable to extinct and extant organisms. The framework allows quantitative measurement of within-assemblage ecological diversity (life habit richness and disparity, measured by mean Euclidean distance and maximum range). Dynamics of these metrics as a function of alpha species richness are examined in computer simulations modeling the rules of the four hypotheses. These dynamics are then compared with empirical estimates of ecospace occupation in Paleozoic (Cambrian-Devonian) and modern deep-subtidal, soft-substrate assemblages.

Initial simulation models overestimate the diversity of actual assemblages because organisms are constrained in their life habits. Modifications incorporating more realistic assumptions (such as constraining potential life habits according to actual frequencies and adding a parameter controlling the strength of each model's rules) provide better correspondence to actual assemblages.

The null model deserves consideration generally because it replicates many trends predicted by the driven hypotheses. For example, both null and expansion models predict greater disparity in larger assemblages. Overall, this null model provides the best fit to the dynamics of both modern and Paleozoic assemblages, although weakly driven mechanisms can not be ruled out.