Phylogenies and Comparative Biology, Stage II: Testing Causal Hypotheses Derived from Phylogenies with Data from Extant Taxa

Abstract

The revolution in comparative biology that occurred over the past 15 years stemmed from two related developments. In the early 1980s, a number of workers argued that macroevolutionary phenomena can be interpreted only in an explicitly historical context (e.g., Cracraft, 1981; Lauder, 1981; Gould and Vrba, 1982). Shortly thereafter, workers realized that as a result of shared ancestry, species are not statistically independent entities; consequently, statistical analyses of comparative data are invalid unless phylogenetic information is incorporated (e.g., Gittleman, 1981; Ridley, 1983; Felsenstein, 1985; Harvey and Pagel, 1991). The result is that workers in all fields of biology are now aware that phylogenetic information must be incorporated into any comparative study that investigates causal hypotheses (i.e., studies that are not purely descriptive). The number of journal articles that incorporate phylogenies has increased substantially, not only in those journals traditionally devoted to evolutionary issues but also in journals such as Animal Behavior, Ecology, and Development. Nonetheless, the phylogenetic revolution is only half complete. Although workers now commonly use phylogenetic methods to evaluate hypotheses developed using other forms of data, the time has come to invert the process and to use phylogenies as the source of testable hypotheses about evolutionary patterns and processes. Phylogenies are statements not only about relationships among taxa but also about the evolution of characters. These statements in turn may suggest causal hypotheses about why character change should occur in a particular manner. In many cases, hypotheses deduced from phylogenies make predictions that pertain to both historical events and contemporary phenomena. For example, phylogenetic analysis indicated that in ray-finned fishes the hypochordal longitudinalis muscle arose prior to the evolution of an externally symmetrical tail (Lauder, 1989). One hypothesis that could be derived from this observation is that the muscle alters functional capacities of the tail in such a way as to favor a symmetrical tail, which might not have been beneficial previously. Functional studies of the hypochordal longitudinalis in extant fish support this hypothesis by revealing that the muscle qualitatively alters tail function (Lauder, 1989). In a similar manner, testable hypotheses can be derived from phylogenies in many fields of comparative biology as the first four papers in this issue of Systematic Biology attest. The form that such tests take, however, will vary from field to field. Laboratory examination and manipulations will be most appropriate to fields such as developmental biology, functional morphology, and endocrinology, whereas held observations and/or experiments may be more appropriate in ecological or behavioral studies. In addition, measures of genetic variation and constraint can be useful to test some hypotheses (e.g., Futuyma and McCafferty, 1990). In some situations, one can test hypotheses that certain traits are favored by natural selection in particular selective regimes (sensu Baum and Larson, 1991) as predicted based on phylogenetic evidence; multigeneration experiments can test whether selection leads to evolutionary change in the direction predicted by phylogenetic hypotheses. This