Stochastic Simulation and Evolution of Morphology-Towards a Nomothetic Paleontology

Abstract

Raup, D. M. (Department of Geological Sciences, University of Rochester, Rochester, N.Y.) and Gould, S. J. (Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts) 1974. Stochastic simulation and evolution of morphology-towards a nomothetic paleontology. Syst. Zool. 23:305-322.-The order of evolutionary trees has been the traditional argument for the operation of directional causes in macroevolution. We show, in this work, that a similar order can be generated within stochastic systems bounded by minimal biological constraints. Our system generates an evolutionary tree by making random decisions about each lineage in each time interval given preset probabilities for branching, extinction and persistence (Raup, Gould, Schopf, and Simberloff, 1973). Morphology is determined in an independent and equally stochastic manner. Using ten hypothetical characters, the beginning lineage is given an all zero morphology. At each branching point, each character may change by one unit (in a positive or negative direction) according to preset probabilities for positive change, negative change, and no change. Our simulations display most of the ordered features generally associated with uni-directional selection: coherence of monophyletic groups and incomplete filling of morphological space; regular unfolding of morphology (as seen in strong correspondence between phenetic and cladistic taxonomies); marked evolutionary trends; strong correlation among characters; large variation in rates of evolution; and specialization of derived forms. We attribute much of this order to abstract topological properties of the tree itself and urge that the data for inferences about directional causes be sought elsewhere (in functional morphology, for example). We suggest, with caution, that undirected selection may be the rule rather than the exception in nature, if a temporal unit of sufficient duration be used as the yardstick of measurement. [Simulation; evolution.] Modern paleontology has retained a vestige of idealistic morphology in its traditional argument for the role of directional causes in macroevolution. The presence of order in the results of evolution has been taken uncritically as definite evidence for the production of such order by directed causes. The results of evolution are the phylogenetic trees that paleontologists love to draw. The marks of order include evolutionary trends, correlation between characters, and coherence of taxonomic groups. The postulated cause, for the last thirty years at least, has invariably been uni-directional selection. We have come to doubt that the formal pattern of change is an adequate argument for directed causes. We explore our doubts in this paper by studying the order that can arise in random systems of change, bounded only by the conventional assumptions of monophyly, continuity, and equilibrium. In an earlier paper, we reported the results of a stochastic simulation of phylogeny (Raup, Gould,, Schopf, and Simberloff, 1973). We generated phylogenetic trees by a random process and interpreted them in a conventional paleontological manner. We studied patterns of origination and extinction of higher taxa and the evolution of taxonomic diversity. Throughout the work, we did not attempt to replicate nature but rather to identify those aspects of actual phylogenies which behave as stochastic or random variables. (We did this, in part, to establish criteria of subtraction for the identification of phenomena in actual phylogenies that cannot be simulated by random processes.) The study was entirely cladistic; it dealt only with patterns of branching and extinction of lineages or