[Taxacom] New systematics book

[Curtis Clark]

First I want to say that, from all the agreement on this list that 
species are not entities in nature, it seems that pattern cladistics has 
won, and those of us who studied (and perhaps still study) speciation 
could have done no worse by studying unicorns. All the rest of what I 
write is based on the assumption that species are the result of an 
evolutionary process, and to that extent (at least) are real. I surmise 
that you agree.

I accept paraphyletic species, and in fact my own studies suggested 
several examples. What I have never understood is how the existence of 
paraphyletic species provides a general justification for paraphyletic 
higher taxa. If the pattern of character distribution in extant lineages 
showed, say, four lineages that share a synapomorphy, three of them have 
distinct autapomorphies, and the fourth has no autapomorphies, that 
could be suggestive of ancient peripatric speciation. And it would be a 
clear case for a polytomy in a cladogram. But paraphyletic taxa like 
Reptilia (sensu antiquo)  and Dicotyledonae are most often the result of 
gaps in knowledge; it's hard for me to imagine the existence of 
peripatric speciation justifying them.

On 2013-09-08 1:35 PM, Richard Zander wrote:
> " ... parsimony and other optimality analyses only work with accuracy
> with pseudoextinction (dying ancestor, two new species). The correct
> optimality for branching analysis is that which makes the shortest (most
> likely, most credible) tree given identified surviving ancestors and
> their daughter taxa based on other information. This is merely, MERELY
> less precise, correctly less precise and therefore with more uncertainty
> about evolutionary relationships.

This is a methodological issue, not a conceptual one. There are many 
other problems with parsimony as well (including long branch attraction 
for molecular data). I don't see how this supports recognition of 
paraphyletic groups.

One of the reasons that the results of the Angiosperm Phylogeny Group 
are so widely accepted is that the groups are things we were considering 
anyway, but didn't quite have the evidence for. In a previous post, I 
gave some examples from Scrophulariaceae s.l.; plant taxonomists have 
known for years that the family was heterogeneous, and now we have a few 
more lines of evidence to convince us of that. Just because some people 
make classifications from bootstrapped molecular parsimony trees with no 
other input doesn't invalidate cladistics.

> "...  molecular "lineages" are molecular strains, and ancestral taxa in
> stasis generate lots of molecular strains before and after generation of
> one or more daughter species. The DNA continues to mutate in isolation
> while expressed traits remain in stasis for millions of years. Molecular
> phylogenetics gives precise sister group analyses of extant strains, not
> taxa. Evidence? Molecular paraphyly. What about extinct or unsampled
> strains? They could have diverged molecularly long before or after any
> extant strain."

This is again methodological. The lack of congruence of many gene trees 
is well-known, as are the evolutionary processes resulting in the lack 
of congruence. Again, no basis here for general recognition of 
paraphyletic higher taxa.

Within the eukaryotes, the general pattern of groups within groups is 
pervasive. Peripatric speciation, hybrid speciation, and lateral gene 
transfer are all fascinating processes that certainly muddy the waters, 
but there are a lot of "just-so stories" lurking in Bessey's Cactus and 
its descendents. Perhaps your book has a methodological framework to 
address that (I presented a method for defining grades a year or two 
ago, but you neither challenged nor agreed).

And as far as "phylogenetic jargon has taken a life of its own," it's 
perhaps not a good idea to redefine terms like monophyly yet again.

-- 
Curtis Clark        http://www.csupomona.edu/~jcclark
Biological Sciences                   +1 909 869 4140
Cal Poly Pomona, Pomona CA 91768