|E. O. Wiley (2008), Scholarpedia, 3(12):6299.||doi:10.4249/scholarpedia.6299||revision #91654 [link to/cite this article]|
Phylogenetic systematics is the discipline of reconstructing the common ancestry relationships of organisms and constructing formal taxonomic classifications that are logically consistent with these relationships.
Phylogenetic systematics grew out of the synthesis of the German entomologist Willi Hennig (1913-1976), whose books on the subject published in 1950 in German and modified into English and Spanish in 1966. Hennig presented on three ideas. First, the most basic relationships among organisms were genealogical/common ancestry relationships and not similarity relationships. Second, only certain homologous characters could confirm the basic hypothesis that two species are more closely related to each other than to a third species. Third, natural groups of species are those that include an ancestral species and all descendants of that species. These ideas were not new, as Hennig acknowledged, but his synthesis presented a new paradigm to evolutionary biology.
The idea that the most basic kinds of relationships are genealogical is generally traced to the emergence of the evolutionary paradigm. The tree of life is one continuous stream of genealogical descent punctuated by division of the stream that appear as separate branches, as reflected in the one illustration appearing in Darwin's Origin of Species. However, the pre-evolutionary idea of grouping organisms on the basis of similarity rather than genealogy was well-entrenched in taxonomic thought, even when scientists accepted evolution as the basis for observed similarity. Two modern movements asserted that either overall similarity was the criterion for grouping (phenetics) or that one should use a combination of genealogy and similarity to group ("traditional" evolutionary taxonomists). The problem with these ideas is that it is impossible to argue that one measure of similarity is superior to any other. Hennig, building on the work of others, provided a way of sorting out those similarities.
The idea that only certain kinds of homologous characters could be used to demonstrate common ancestry is derived from the observation that evolutionary innovations (potential new homologies) arise at different times and in different lineages. Fins evolved before legs and flowers after phloem. Thus, some homologies are shared because of relatively recent common ancestry while others are shared because of more ancient common ancestry. While all homologies show common ancestry relationships if the entire tree of life is considered, only some show common ancestry relationships if only part of the tree is considered. Phylogenetic analysis is used to argue the relative antiquity of homologies and whether or not two similar characters are, in fact, homologous at all.
The first step in phylogenetic analysis is to define the problem to be solved. Given that all organisms are ultimately related, we wish to ask questions only about some subgroup of the entire tree of life. Any two species are always related, given a single origin of life, so the most basic problem is one that involves three species. Consider three species: a tuna, a human and a lizard. There are only four ways these three species can be related: Figure 1.
The second step is to analyze the similarities and differences we observe among these species. Tunas have fins, humans and lizards have legs. In this respect, humans and lizards are similar. Detailed study of the structure, development and even genetics of fins and legs indicates that fins and legs are homologous and the details of human and lizard legs indicate that the legs are homologous. That is: we conclude that, at least for the moment, fins and legs most likely have a single evolutionary origin and that the legs also have a single evolutionary origin.
The third step is to determine which came first, the legs or the fins. If the common ancestor of tunas, lizards and humans had legs, then we cannot argue that humans are more closely related to lizards than to tunas because the common ancestor of tunas also had legs. But, if the common ancestor of tunas, humans and lizards had fins, then legs are a relatively recent evolutionary innovation and the property of having legs would indicate that humans and lizards are more closely related to each other than either is to tunas. How do we discriminate between these alternative hypotheses?
We discriminate by looking outside the group and see what other organisms have, especially closely related organisms outside the confines of the group. In this case, there is no use in looking at birds, hippos, turtles, or basses, they are all inside the group. What we need is a group that is closely related but no in the group. Sharks are a good candidate. Sharks have fins. Using sharks we can conclude that the ancestor of tunas, humans and lizards had fins; that fins are the primitive homology and that legs are the derived homology. As the derived homology, legs are valid confirming evidence that humans and lizards are more closely related to each other than either is to a tuna. We cannot conclude that basses are more closely related to tunas than either is to a lizard because both fishes have fins; nor can we conclude that tunas and basses are closely related to sharks. Why? The common ancestor or all these species (basses, tunas, sharks, lizards and humans) had fins. But, we can argue that all these species had a common ancestor because hagfishes, lampreys, amphioxus, and tunicates do not have fins. Fins (pectoral and pelvic fins) are an evolutionary innovation that arose in the ancestor of all jawed vertebrates.
All phylogenetic analysis is based on the reasoning presented above. All homologies are derived at some level of the tree of life. They confirm relationships at the level they are derived and once derived provide no further confirmation. Paired fins confirm the hypothesis that all jawed vertebrates share a common ancestral species but do not confirm that sharks are more closely related to tunas than tunas are to lizards.
In any one section of the tree homologies appear in two varieties. In some cases, every species of the group have the homology or something derived from that homologue (all jawed vertebrates have paired appendages except some groups like snakes where the adults lost them). In some cases, they have different, but homologous characteristics (some vertebrates have fins, others have legs). Plesiomorphic characters do not show relationships within such a group because they already show relationships of a larger group. Fins confirm that jawed vertebrates are all related, legs show that tetrapod vertebrates form a group within jawed vertebrates. The hierarchy of genealogical evolution is reflected in the hierarchy of character modification (Figure 2).
Of course, not all similarities turn out to be homologous similarities. Sometimes very similar characters are evolved independently in different ancestors. There are two ways to sort out homologies and non-homologous (homoplasous, convergent) similarities. One way is to look at the characters in detail. For example, one might initially think that the fangs of marsupial lions and African lions are homologous. But, the fangs are actually different teeth. The other way is to test the characters using other characters. Homeothermy in birds and mammals is considered a convergent similarity because all the other characters that have been analyzed indicate that the common ancestor of birds and mammals, a common ancestor that also gave rise to turtles, lizards, snakes, and others, was not warm-blooded.
Phylogenetic analysis of relatively simple data, such as that shown in Figure 2. , can be carried out by hand. But, as data accumulate it becomes harder to insure that all possibilities are considered. Thus modern phylogenetic analysis uses computer algorithms that sort through and fit characters to trees using optimality criteria. There are two common criteria. Maximum parsimony adopts the criterion that the optimum tree is the one that explains all the data in the shortest number of evolutionary steps. Likelihood analysis is a statistical approach that using one or more specific models of how characters might change and uses these models to assess the likelihood of observing the data and model given a particular tree topology. Two different tree topologies can be compared to see if one is statistically better than the other. No such statistical procedure is available for parsimony analysis, but both approaches yield similar results over a relatively wide range of evolutionary models. Yet another approach is Bayesian analysis. It lacks an optimality criterion, but uses likelihood functions to assess the likelihood of the tree given the data and model.
Phylogenetic classification is the third and final component of the Phylogenetic system. Hennig suggested that the most general (and thus most useful) classifications would be those originally advocated by Darwin: strict genealogical classifications. Such classifications would contain groups only comprised of groups that descended from a single common ancestral species and all descendants of that species. Groups that contain an ancestral species and all its descendants are termed monophyletic groups. Hennig contrasted monophyletic groups with two other kinds of groups. Paraphyletic groups contain an ancestral species and some, but no all, of its descendants. For example, the pre-Hennigian concept of Reptilia included lizards, snakes, turtles, crocodiles and dinosaurs, but excluded birds. Group that do not contain the common ancestor are termed polyphyletic groups.
No one advocated polyphyletic groups. The controversy that ensued was over the nature of paraphyletic groups. Before Hennig, such groups were considered to be monophyletic groups. But to Hennig, such "incomplete" groups were as artificial as polyphyletic groups. The problem was that there were (and continue to be) many well-loved and familiar paraphyletic groups. Reptilia without Aves is paraphyletic. Pongidae without humans is paraphyletic. Lungfishes are more closely related to cows than to tunas. But, paraphyletic groups, much beloved and traditional as they might be, face an insurmountable problem: their inclusion into formal classifications renders the classifications logically inconsistent with the phylogenies of the organisms (a point made even before the general acceptance of Hennig's methods by the philosopher David Hull). As a result, modern classifications do not recognize such groups. In a modern classification, Reptilia include, not exclude, birds.
- Brooks, D. R. and D. A. McLennan. 2002. The Nature of Diversity: An Evolutionary Voyage of Discovery. University of Chicago press.
- Hennig, W. 1966. Phylogenetic Systematics. University of Illinois press, Urbana.
- Judd, W. S., C. S. Campbell, E. A. Kellogg, and P. F. Stevens. Plant Systematics: A Phylogenetic Approach. Sinauer.
- Wiley, E. O. 1981. Phylogenetics. The Theory and Practice of Phylogenetic Systematics. Wiley-Interscience, NY
- Wiley, E. O., D. Siegel-Causey, D. R. Brooks, and V. A. Funk. 1991. The Compleat Cladist, A Primer of Phylogenetic Systematics. Spec. Publ., Museum of Natural History, University of Kansas.