Friday, April 1, 2011

Book Review: Foundations of Systematics and Biogeography.

Ebach and Williams "new" book has been out for several years now, and while Platnick, Brower, and Amorim previously weighed in, none of these are as critical and scathing as Ferris's recent review in Cladistics.

The book does, though, contain things we did not know—indeed, there are things here that no one has ever known. Among these revelations are that cephalaspids are placoderms; that Haeckel was responsible for grouping by overall similarity whereas Cuvier and Huxley had nothing to do with it; that Sokal and Sneath (1963) were primarily interested in phylogeny; that Nelson never accepted transformation; that Hennig used only congruent characters, because Sokal and Sneath said so; that paraphyly has no connection with symplesiomorphy; that DNA sequences have no hierarchic structure—unless used by Patterson; that parsimony is limited to binary characters, is closely linked to grouping by overall similarity, and does not draw conclusions of homoplasy; that reversals do not really happen and would be plesiomorphies if they did; that all un3ts data are phenetic—unless used by Patterson; that 3ta and ppa are not methods, being inspection of the data instead; that transformation is a myth; and that transformation is nonetheless distortion. A reader would have to be rather poorly informed to fall for this nonsense, which is to say that the book has obviously been designed to victimize just such readers. Ignorance is strength. This brings us back to Platnick's (2009, p. 281) central comment:

It should be especially useful for students.

Like Williams and Ebach, he has realized that tricking the least experienced students is the only chance that 3ta has left. No doubt publishers relish such recommendations. If Platnick’s endorsement were believed, Springer would get their $90 a head regardless of the damage done to students.


I haven't yet had time to purchase a copy and read it, so I can't really weigh in from my perspective. Generally, I trust the evidence Ferris presents in his review. The main conflict centers around the use of three taxa analysis (3ta), and whether or not it is a useful method for phylogenetic inference. This old method takes three taxa from a set to be analyzed, decides which two are more closely related to each other than to the third, and puts them back into the mix, rinse, repeat, until all relationships are resolved. This seems to me to be a wholly inefficient way of parsimoniously finding a summary of relationships, and according to Ferris's review, may be more likely to create paraphyletic groupings than path finding the shortest trees by heuristics (approximation) and collapsing to a consensus. Consider if you had 30 species to search, which is millions of combinations. This is not the same as simply trying to resolve the relationship between three taxa alone. The simplicity of such an arangement means there are only three hypotheses (topologies), of which one is correct. An example would be simply trying to find the relationship of Ephemeroptera, Odonata and Neoptera. Relationships within each of those groups or outside are irrelevant to the question: are Ephemeroptera and Odonata more closely related to each other than either is to the Neoptera? Another issue I find is the emphasis that a so called "reversal" of a character could never be considered a synapomorphy. To which I ask, at what level?

Farris's comments makes the whole of Foundations seem to be a strange reactionary piece against traditional pattern cladistics, whereas traditional pattern cladistics is seen as reactionary by many molecular biologists today. I am not saying I'm going to remove Urhomology from my blogroll; I still value many of the statements Ebach and Williams have made there. It was while reading through the backposts I was inspired to write the statement "Homology is the key to the heart of biology", which continues to be a motto for me in regards to the importance of homology (synapomorphy) in figuring out the history of life and the pattern of evolution. I appreciate their rejection of paraphyly and their use of the term phylophenetics when discussing molecular phylogenetics as it is done currently. Even if the authors dwell on false or bad premises it does not carry that those ideas could not be instrumentally true in regards to the right premises.

When I finally purchase and read Foundations of Systematics and Biogeography, I'll let you know if my opinions change.

1 comment:

David Williams said...

I've only just stumbled upon this. Your statement "This old method takes three taxa from a set to be analyzed, decides which two are more closely related to each other than to the third, and puts them back into the mix, rinse, repeat, until all relationships are resolved" is not an accurate summary of three-item. Maybe this is not the right place to do provide a more accurate summary but just to flag up the fact that yours is somewhat misleading.