Ancestors, Descendants, Sister Groups and Testing of Phylogenetic Hypotheses

Abstract

Szalay, F. S. (Department of Anthropology, Hunter College, City University of New York, 695 Park Avenue, New York, New York 10021; Department of Vertebrate Paleontology, American Museum of Natural History, New York, New York 10024) 1977. Ancestors, descendants, sister groups and the testing of phylogenetic hypotheses. Syst. Zool. 26:12-18. -It is argued that the practice of expressing all phylogenetic relationships as cladistic relationships (i.e. splittings, those expressed only in the form of dichotomous sister group relationships) gives an incomplete and therefore potentially misleading view of evolutionary history. Consequently such schemes tend to supply hypotheses that fail to mirror evolutionary descent. Application of shared derived characters to set up phylogenetic hypotheses is a mere consequence of establishing morphocline polarities. Therefore, the of a phylogenetic hypothesis lies in the soundness of biological (developmental, functional, etc.) investigations of character clines. challenging issue then is the testability of morphocline polarity hypotheses, rather than competing phylogenetic schemes. Age of any phenon is one of its attributes, its fourth dimension. In addition to those same considerations as are employed in dealing with sister group hypotheses (shared-derived characters), ancestordescendant hypotheses may also be tested by the temporal position of the taxon. According to the Popperian view of criteria, postulated ancestor-descendant relationships may be falsified by biostratigraphic evidence, and therefore such hypotheses are as scientific as those of a sister-group nature. Phylogenetic trees are preferred over cladograms because the former always contain more information since they can express both ancestor-descendant relationships (anagenesis) as well as sister group relationships (cladogenesis). A brief synopsis is offered for the procedure to establish historical relationships of taxa. [Phylogeny reconstruction; morphocline polarity hypotheses; ancestor-descendant relationships.] The relative age of different characters should in all cases be a prime object of research. This historical method (although open to many pitfalls) when judiciously applied seems more likely to lead to lasting phylogenetic results than the time-honored method of setting down all the resemblances and differences between two animals, without furtlher analysis, and then striking a balance at the end. The greatest stumbling blocks of the phylogenist lie: first in the difficulty of distinguishing between primitive and specialized characters, secondly in the tendency to assume relationship between two given forms on the basis of resemblances that may have been brought about by either parallel or convergent evolution. W. K. Gregory (1910) A genealogy or phylogeny involves both sister-group and ancester-descendant relationships. Increasingly, however, in what has become known as cladistic analysis, expression of all hypothesized phylogenetic relationships has become restricted to sister-group relationships even in instances when the fossil record is excellent (see for example, Eldredge and Tattersall, 1975). As evolution embraces the phenomenon of splitting as well as ancestor-descendant relationships, it is useful to briefly examine the exclusive practice of employing the concept of dichotomies without ancestordescendant relationships, the use of cladograms, and various claims as towhat really constitutes the test of a phylogenetic hypothesis. SISTER GROUPS AND ANCESTOR-DESCENDANT