We analyzed planktic foraminiferal evolutionary data using techniques of nonlinear dynamics, a methodology new to paleontology. The data set comprises 196 extinction and speciation horizons derived from biostratigraphic ranges of 662 reliably defined species. Both extinction and speciation data sets are well characterized by power-law models. However, return maps and a predictor technique indicate that the extinction data are more highly deterministic than speciation data. We interpret the analysis, particularly extinction data, to be consistent with planktic foraminiferal evolution being organized, and not randomly driven. Our results do not preclude periodic large extinction events driven by external forces as predetermined by another system (e.g., large-body impact), or internally driven extinction processes, where spontaneously derived interdependencies cascade through the ecosystem, or some combination thereof. Our data support a model whereby the internal organization of an ecosystem regulates the response to changes in a deterministic manner, the relative scales of disturbances and extinctions depending on the degree of interdependency within the system. Thus any contention that species interactions are not sufficiently intense to generate mass extinctions can be dismissed. Random walks generated by genetic drift and the transitory nature of n-dimensional niche space may explain why speciation is less deterministic than extinction.