Examine macroevolutionary concepts carefully Nicholas Matzke, Grad student, UC Berkeley 29 August 2007 Well, since it is clear that this paper will be on every ID/creationist blog on the planet in under 12 hours, I might as well put in my 2 cents early. I will say at the outset that I have immense respect for Eugene Koonin and his contributions to numerous fields, and criticize his work with some trepidation. However, I think with this paper he has unfortunately tumbled into a series of mistakes that have repeatedly afflicted those trying to understand macroevolution without taking sufficient care in examining the concepts they are relying upon.I will enumerate some of them, in no particular order:1. Treating Linnaean categories as "real", and thinking about them in a typological manner rather than phylogeneticallyThis is most clearly exposed in Koonin's discussion of the Cambrian phyla, wherein it is innocently assumed that there really are things called "phyla" that are non-arbitrarily defined, and it is also assumed that the series of characters that are associated with a particular "phylum" always occur together and "emerged suddenly" with no evidence of transitions. This was a view found in Gould's Wonderful Life and numerous other analyses of the "Cambrian Explosion", but they are all hamstrung by uncritical acceptance of the Linnaean system's "phylum" category.In reality, as the cladistics-informed paleontologists have been trying to point out to everyone for some time, apparently still not completely successfully, "phyla" cannot be given a non-circular definition. We can identify the crown groups (groups defined by a monophyletic set of living organisms, their common ancestor, and the descendants of that ancestor), but assigning them a particular rank is basically an arbitrary procedure. Studies of early Cambrian fossils (read Graham Budd) show numerous fossils with some, but not all, of the characters that "define" the living "phyla." Should we "lump" them into the living phyla, thus obscuring their transitional nature? Should we put them in their own phyla, because obviously the extinct sister group of a living phylum should have equal rank? But this often erects an entire phylum on nothing but a few minor character differences. The problem is unsolvable, because fundamentally Linnaean categories are typological and evolutionary transitions are not.(Another way to see the absurdity of "phyla" is to compare animal "phyla" with protozoan "phyla" or bacterial "phyla". Phylogenetically, all animals (a "kingdom") with all their spectacular diversity are just one subgroup of protozoa)Paleontologists can and have put the Cambrian animals, e.g. arthropods and close fossil relatives, into cladograms, and the order of step-by-step acquisition of characters -- i.e., a gradual transition ("gradual" = "graduated", =step-by-step, not "smooth constant-rate-ism") has been reconstructed. These prove that the origin of characteristics that we now associated with the arthropod "phylum" did not all originate at once, instead they were assembled piecemeal. The transitional grades do exist in this case, it's just that Koonin is unaware of them. The appearance of suddenness is based on staring at the living groups (or equivalently, staring at their molecular trees), treating them typologically, and then mistakenly assuming that the lack of living transitional forms means that there never were any. When you have a decent fossil record, this provides the stem groups below the crown, and inevitably (cambrian phyla, birds, whales, mammals, etc.) this shows that apparently puzzling "big bang" transitions occurred in a long series of stages. For more on the "phylum" problem and the importance of considering stem groups:Down with phyla!http://www.pandasthumb.org/pt-archives/000969.htmlDavid Fitch and Walter Sudhaus, “One small step for worms, one giant leap for ‘Bauplan?’” Evolution & Development 4:4, 243-246. http://www.pandasthumb.org/archives/2005/08/down_with_phyla.htmlhttp://www.pandasthumb.org/archives/2007/04/meet_orthrozanc.htmlGraham Budd (2001). “Climbing life’s tree.” Nature 412, 487.Graham Budd (2003). "The Cambrian Fossil Record and the Origin of the Phyla." Integrative and Comparative Biology, 43(1), pp. 157–165Obviously, something happened early on in the Cambrian, and several triggers and causes have been proposed. But once it got going, fundamentally it was just another adaptive radiation, probably more dramatic than later ones because many niches were being occupied for the first time.More quickly:2. Another problem is the failure to consider fundamental features of phylogenies as they approach the base of the tree. It is simply an unavoidable, automatic, intrinsic feature of phylogenies based on crown groups that you will have less and less information about events as you trace them back. And it is inevitable, automatic, and intrinsic that the crown groups will seem more and more distinct -- the modern organisms, which is what we have to look at, are actually less related, so they are more distinct! Assuming common ancestry of life on an arbitrary life-supporting planet, this would happen *no matter what* actually survived to be discovered by observers 4 billion years later. No matter what, eventually the phylogeny would trace back to a few fundamental groups, and these groups would appear to be the most distinct, the least information would be available about their origins, and they would provide the most fodder for speculation that something "unusual" must have gone on.Another issue is that signal-to-noise declines as you go back billions of years, so a prosaic null hypothesis is just "we don't have good enough data yet", but Koonin is aware of this so I won't continue on this point. Unfortunately this paper does not attempt to test the "big bang model" versus the prosaic null hypothesis of "not enough signal."(I also think it would have been helpful for Koonin to look at the improvements in the resolution of our phylogenies, not just the remaining problem areas -- by any reasonable standard there has been immense progress in the last 20 years, probably more progress than was thought possible 40 years ago. Cavalier-Smith's recent work on eukaryote phylogeny is an example of what I am thinking of.)3. Even without fossils, it is utterly clear that a great many "origin of complexity" events occurred by what I guess Koonin would think are boring "orthodox" processes. The origin of the adaptive immune system is a good example. Phylogenies show that many of the "key" features of adaptive immunity existed before adaptive immunity itself emerged in gnathostome ancestors, and the "key" event of the transposon insertion was followed by round after round of duplication and divergence to produce the multitudes of V, D, J segments. The end result is fantastically complex, but the causal process is pretty simple and prosaic once it is understood. (A recent paper deflating the common "big bang" rhetoric surrounding the origin of the immune system was the following:Klein, J. and Nikolaidis, N. (2005). "The descent of the antibody-based immune system by gradual evolution." Proceedings of the National Academy of Sciences 102(1): 169-174.See this link for a history of the topic:http://www2.ncseweb.org/kvd/exhibits/immune/immune_evo_annotated_bib.html )Similarity, there is plenty of evidence for duplication and divergence of genes having occurred e.g. in the ancestors of crown eukaryotes, in the ancestors of the LUCA (whatever the nature of this was, exactly). This is direct evidence of simpler precursors and step-by-step acquistion of complexity.4. A more specific issue is that one of these major transitions, the origin of multicellularity, seems to be unambiguously as gradual as you like, having occurred probably dozens of lines, often phylogenetically recently. Again the key processes are simple and stepwise. With animals we can reconstruct a set of transitional grades going back through sponges and choanoflagellates.I think the reason that this particular major transition inspires less talk of unusual evolutionary processes is that the character of interest, multicellularity, is so widely known to be a common evolutionary occurrence. For some reason, though, other characters in other situations get rarified and attached to typological definitions of a a crown group to create the illusion of unusually "big" changes. I think this has happened with many eukaryote characters, when in fact many of the "key" features of eukaryotes can be found to have originated independently in one or more prokaryotic groups.In summary, I think Koonin should give a little credit where credit is due to gradual, stepwise evolution, and not try to argue that Darwin's eye discussion is all that orthodoxy has going for it. Being revolutionary and unorthodox is much more fun, of course, but sometimes you've got to wonder how many hangovers (i.e., creationist quote-mining and general confusion over the status of evolution outside of the specialist community, and needless wrangling within the specialist community) could be avoided if scientists would exercise just a little caution during the party (i.e., spending a little time soberly comparing their revolutionary ideas with more prosaic explanations). Competing interests Until this week I worked at the National Center for Science Education, where we oppose the ID/creationists and develop a finely-tuned sense of the sorts of things they will pluck from the literature and desperately portray as evidence that they aren't completely nuts. However, I am well aware that telling scientists to censor themselves to avoid giving creationists talking points is a non-starter, so hopefully my comments came out as being substantive rather than just the boring voice of orthodoxy.