[Taxacom] New systematics book

Richard Zander Richard.Zander at mobot.org
Mon Sep 9 12:44:19 CDT 2013

I appreciate the response, pro, con, and incensed. Okay, here is the
shortest way I can describe what passes for optimality in phylogenetics.


Take two taxa or clades (at the end of a larger cladogram with other
taxa or clades), one characterized by advanced traits xy and the other
by advanced traits xz. Phylogenetics assumes that parsimoniously the
shared ancestor, now extinct through pseudoextinction (anagenetic change
into another taxon), had the plesiomorphic trait x, then one daughter
taxon differentiated with y and the other with z. That is three (3)


Okay. BUT suppose the ancestral taxon did not go extinct but survived
and generated a daughter taxon by peripatric evolution. With the same
data set, the ancestral taxon had traits xy, and the daughter taxon had
trait z and -y (reversal). That is four (4) steps.


In phylogenetics, optimality is always BOTH shortest tree and simplest
model (i.e., universal pseudoextinction). The argument is that you do
get the shortest tree with both assumptions. Of course the parsimony or
other optimality part is okay but the simplest model is nonsense, since
this sort of argument would throw out that complicated theory of
relativity in favor of the simpler Newtonian mechanics. Occam's Razor
does not work that way since the Razor deals multiple explanations for
one model.


My suggestion is that the best model for any optimality (morphology or
molecular) analysis is generating a mixture of pseudoextinction and
peripatric evolution for a group, THEN find maximum parsimony,
likelihood or credible interval for that evolutionarily composite model.


The old complaint is that detailing such a composite model requires
experience, judgment and reasoned evaluation based on theories of taxon
transformation in nature. (Read: subjective just-so story.) It is in
fact easy to examine a group of classically grouped species and discern
generalized species and highly specialized potential daughter species.
THEN constrain the cladogram. 


Although using universal pseudoextinction makes for ease of generation
of cladograms, the results are in often large part imaginary, that is,
all divergence branches involving a single surviving ancestral taxon are
not real. 




Richard H. Zander

Missouri Botanical Garden, PO Box 299, St. Louis, MO 63166-0299 USA  

Web sites: http://www.mobot.org/plantscience/resbot/ and

Modern Evolutionary Systematics Web site:

UPS and FedExpr -  MBG, 4344 Shaw Blvd, St. Louis 63110 USA


More information about the Taxacom mailing list