Levels of selection and large-scale morphological trends

Large-scale morphological trends involve the concerted evolution over macroevolutionary time-scales of a large number of constituent species. Many mechanisms have been suggested to be involved in generating trends. Here I develop methods for the detection of species selection and apply them to simulated and empirical data. The camerate crinoid calyx undergoes a simplification trend over the group's evolutionary history. Possible general trend mechanisms include: A tendency for descendants to be simpler than their immediate ancestor, termed a bias in the direction of speciation. Species selection, where differential diversification of species sharing particular morphological attributes. If species selection alone generates this trend, species with simpler calyces will, as a whole, have higher net-diversification than species more complex calyces.; I find that species selection alone could not generate the observed trend. The subclade test is used to indirectly indicate the presence of a bias in the direction of speciation, which could account for the overall directionality of the trend. Skewness of subclades indicates that there is inherent directionality, either from anagenetic or cladogenetic means. The fact that species selection occurs with a directional bias in camerates means that the relative frequency of "Wright's Rule" cannot be used as a measure of the relative frequency of species selection. A new approach to studying the shape of taxonomic survivorship curves is developed and presented that shows promise for measuring the relative frequency of species selection.; Many factors shape taxonomic survivorship curves. But the influence of the extinction and speciation rates of constituent species on the shape of genus-level curves has been underexplored. I propose a new model of hierarchical extinction, where extinction of subordinate species is tightly coupled with the extinction of genera. All species extant at the time of genus-level extinction will simultaneously go extinct. The difference between the rate of genus-level and species-level extinction determines if the genus-level survivorship curve is concave-up or log-linear. The hierarchical extinction model, Raup's upward model, and Van Valen's Red Queen's hypothesis of exponential survivorship are simultaneously empirically evaluated against the genus- and species-level survivorship.