Everett C. Olson

Community Evolution and the Origin of Mammals

The evolutionary course from primitive pelycosaurian reptiles through therapsids to mammals can be profitably studied in relationship to modifications of the structure of the communities in which these reptiles existed. For this purpose the community is defined in very board terms. Three types of communities are recognized upon the basis of the nature of the food chain. Each has an important tetrapod component. Early phases of the evolution that culminated in mammals took place in communities that were strongly tied to water by the structure of the food chain. The physiological bases of the development of mammals appear to have been related to this environmental restriction. In successive pulses, however, the pelycosaur—therapsid communities developed terrestrial reptilian herbivores and thereby broke with the water—based food chain. More strictly terrestrial communities developed concurrently, with the insects, which were a food source for the reptiles, as the principal herbivores. From this sort of community came the terrestrial lepidosaurian—archosaurian reptilian radiation. The terrestrial communities so developed came into competition. In this competition the therapsid lines were temporarily unsuccessful, leaving only small, but very mammal—like, representatives in the late Triassic. After a long period with relatively little adaptive radiation, these remnants provided the basis for the radiations of mammals that led to the great successes of the Cenozoic era.

THE EVOLUTION OF A PERMIAN VERTEBRATE CHRONOFAUNA

case, the vertebrates. Interaction of the animal populations that constitute fauna, either direct or indirect, is implied by the association of the animals, although it is not explicit in the definition of a fauna. When the factor of duration through time is introduced, the term fauna is no longer strictly applicable and may lead to confusion. For this reason the term chronofauna 1 is proposed to differentiate faunal units in which time is an important factor from those in which it is not. A chronofauna is defined as a geographically restricted, natural assemblage of interacting animal populations that has maintained its basic structure over a geologically significant period of time. Environment, both biological and physical, plays an important part in most evolutionary studies and interpretations of past environments necessarily are involved in an analysis of a chronofauna. Reconstructions of ancient environments are subject to various difficulties and conclusions are rarely beyond question. Furthermore, there always remain biological and physical factors that cannot be considered. It is possible, nevertheless, to gain information concerning both physical and biological aspects of environment adequate to provide a background for study of many phases of adaptive modifications of animals. Such information becomes especially significant when changes of environment can be related to patterns of stability and modification of the animals of a chronofauna. Once the history of each species of an assemblage is known and its relationships INTRODUCTION

Size-Frequency Distributions in Samples of Extinct Organisms

The problem of the meaning of size-frequency distributions of measurements taken on samples of fossil organisms is considered. The relationships of such distributions to size-frequency distributions of death assemblages of biological populations is explored by the use of a model based on graphic construction. Rates of growth and mortality are the fundamental biological properties that enter into the model. It is concluded that the form of the size-frequency distribution of a measurement taken on a sample of a fossil species is very little dependent upon the size-frequency distribution of this measurement in the death assemblage of the species. Various problems related to the validity of interpretations based on size-frequency distributions of samples of fossils are examined in the light of the conclusion. The need for careful evaluation of the nature of the sample from evidence beyond that supplied by the size-frequency distribution is emphasized. Means of minimizing the difficulties introduced are suggested.

A MATHEMATICAL MODEL APPLIED TO A STUDY OF THE EVOLUTION OF SPECIES

This paper presents a mathematical model that is believed to be applicable to the solution of certain fundamental problems of evolution. Its use is illustrated by the study of a particular problem. Attention is focused upon the nature of morphological change at a low taxonomic level including analyses of the mechanism and modes of these changes. The data that modern experimental and analytical methodology can provide are as yet insufficient for comprehensive answers to all phases of problems of this type and severe restrictions are imposed when, as in-the illustration, investigations are limited to part of a single anatomical system of extinct organisms. Nevertheless, much of interest and significance may be gained by carefully planned exploitation of the data that are available. Certain taxonomic problems have proved especially amenable to solution by statistical procedures but, mostly, results are either supplementary or complementary to those achieved by other methods of study. There are, however, many problems of evolution that cannot be treated adequately by classical approaches or by application of the statistical procedures now commonly used by students whose data are primarily morphological. Among these are problems concerned with the intricate, small morphological modifications that occur in the evolution of subspecies and species. By use of the method outlined here we hope to detect and describe such changes, thus adding to an understanding of the nature of differences separating small taxonomic units, and also, more importantly, to draw inferences concerning factors that have operated to produce these changes. The characters treated in numerical studies, such as the one considered here, are necessarily dimensions. A serious difficulty in many analyses of the nature of change stems from the tendency to treat modifications of each dimension, as if they were independent of other modifications, whereas intimate interdependence actually exists. Much information, on the other hand, is lost if various changing dimensions are treated together in a way that masks the importance of individual contributions to the whole. Most profitable would be a search for some unifying principle under which each changing part can be studied in relation to its effects on other changing parts, whether these be closely related or partially or completely independent. Such a principle, involving the concept of closely related groups of measurements, has been developed and demonstrated in work, as yet unpublished, by the junior author. We may hope by its use as demonstrated here to approach a little more closely to an organismic basis for analysis of change in evolution, and to provide more definite bases for determining how and why particular changes came about, than has been possible in the past. The problem considered here concerns three closely related species of early Permian reptiles of the Family Captorhinidae, whose essentially contemporaneous existence implies a not remote common ancestor. The acceptance of this implication is the basis for inquiry concerning changes that have probably taken place in the evolution of descendant species, although the ancestral stock is unknown. Mr. William Kruskal, a member of the Committee on Statistics of the University of Chicago, has given much time

The Statistical Stability of Quantitative Properties as a Fundamental Criterion for the Study of Environments

Studies of contemporary environments of sedimentation involve important properties which cannot be measured in the sedimentary rocks. Other properties of contemporary environments of sedimentation can be measured in the sedimentary rocks also, but change diagenetically with time. A third suite of properties remains relatively stable through time. This paper discusses criteria, based on statistical stability of descriptive parameters through time, for treatment of these three classes of measurements for any given property.