(Continued from the previous page)
Nonconcordant predicted phylogenies suggest stabilization processes are far more common than has generally been realized. A phylogeny is a description, typically in the form of a tree of relationships, indicating the evolutionary history of a group of forms. Darwin thought the system of taxonomic classification for any given group of forms could and should reflect the evolutionary history of those forms. He considered this a practical and laudable goal because
The idea that taxonomic classification can, does, and should reflect evolutionary history (i.e., actual phylogeny) is widely considered to be one of Darwin’s most important insights and constitutes what is known as the “cladistic” approach to systematics (a clade is a hypothesized group consisting of a single common ancestral form and its descendant forms, usually thought of as being produced by divergence). Cladists assert that, for any given set of taxa, it is possible to construct accurate trees of descent ("phylogenetic trees”) indicating mutual relationships in terms of the time since any given pair in the set shared a common ancestor.
However, such predicted phylogenies are always hypothetical. They are constructed under the assumption that organisms sharing more traits are more closely related. Conversely, the time since the common ancestor is assumed to be greater for organisms that hold fewer traits in common. Before the advent of modern biotechnology, the traits compared in such studies were ordinary physical features (tooth shape, presence/absence of hooves, number of vertebrae, etc.) and sometimes, also, behavioral traits.
But in recent years the comparison of genetic traits has become more common. For example, the chemical structure of the gene for the enzyme amylase might be compared in various types of organisms. Those in which the gene structure (as measured by various biochemical techniques) is more similar would be assigned to closer branches of the resulting predicted evolutionary tree (predicted phylogeny). On the basis of a second gene, for example the gene for elastin, a second tree of relationships could be constructed, which might, or might not, be the same as the predicted phylogeny based on amylase structure.
Many biologists believe concordance is generally observed between independent gene trees. In their minds, this supposed fact lends strong support to the idea that evolutionary history should be represented as a strictly diverging tree. But, in reality, nothing is more common than disputes over who has published the correct phylogeny. As Whitfield (2007: 248) comments, "different genes from the same set of organisms often predict different trees."
Admittedly, in the case of such disputes there is usually the expectation (given enough time and research, and given the right assumptions and the correct set of traits) that the “correct” tree will eventually be revealed. But in reality, more data does not seem to resolve such disputes. As Milner (1993: 84) points out,
Such discrepancies affect classification even at the highest level—distinct gene trees for archaebacteria, bacteria, and eukaryotes are not concordant. Another example is Hedges' (1999) complete rearrangement, based on more recent molecular data, of the previously accepted phylogeny for Class Reptilia. In studies of a wide variety of other taxonomic groups, a similar lack of concordance has been found. Minelli (1993) cites a profusion of cases in which ongoing controversies are fueled by non-concordant data.
Some researchers have simply given up. After an extensive inquiry, in which she attempted to place the various major invertebrate groups into a treelike scheme of relationships, Willmer (1990) reached the conclusion that it would be impossible to specify any single tree in any way consistent with all the data (recall that 98 percent of all animal forms treated as species are invertebrates). It is well known that the same is true of angiosperms (flowering plants). The vast majority of living plants are angiosperms. As Syvanen (1994: 252) points out, the problem has always been that
It would be impossible to list and consider every existing category of organism, but it may be worth mentioning that Class Pisces (fish), a large, heavily researched, and diverse category, is also in taxonomic disarray. Thus, in the introduction to their Encyclopedia of Aquatic Life (1997), Banister and Campbell acknowledge
If the goal were merely to bore the reader, such cases could be listed almost ad infinitum. It seems sufficient, however, to say that a vast amount of research has failed to demonstrate the existence of a single treelike hierarchy. Such findings prompted Panchen (1992: 243) to assert that the very “existence of a hierarchy embracing all living things is in doubt.” Charles Heiser (1966: 31) put the problem in humorous allegorical form:
The father goes on to tell his son that the fearsome castle is fraught with many perils. But, he says, he who rescues the fair princesses will not only have them as his wives, but also possess their kingdom and all its riches. The son sets forth, and one by one his brothers follow. They take various tools—compound microscopes, scalpels, computers, and pipettes—to aid them in their quests. But, sadly, not one ever returns.
Heiser, who was writing more than forty years ago, says the old taxonomist had one last son, Daniel Niell Alonozo, who "has also left home, and there are some it is said who expect D. N. A., as he is called, to conquer all." However, we see today that data derived from the study of this much-vaunted DNA has failed to broach the dreaded castle's ramparts. In fact, in recent years automated robots and computers, which analyze DNA without human aid, have produced mass quantities of genetic data. This data has put the need for doubt beyond doubt. Thus, Doolittle (2000: 95) commiserates with his fellow biologists on the shortcomings of a discredited hypothesis—the idea of a tree of life:
One obvious hypothesis accounting for this observed lack of concordance in predicted phylogenies, documented so laboriously by so many studies, is that natural hybridization and the production of new forms through stabilization processes is rampant. If the production of new types of organisms through such processes is a widespread phenomenon, as stabilization theory predicts, then the topology of evolutionary relationships would be expected to be largely reticulate (weblike, not treelike) and distinct genes (or distinct sets of traits) would be expected often to yield different trees because, under such circumstances many forms would receive their traits from two different parental forms. In other words, there would be a lack of concordance among trees based on different data sets. And, as we have seen, this is exactly the state of affairs that has been observed. In fact, in many cases we should expect non-concordance. For example, in the case of angiosperms we have long known hybridization is an everyday occurrence. NEXT PAGE >>