[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: something's wrong here: Qianosuchus phylogeny



----- Original Message ----- From: "Jaime A. Headden" <qilongia@yahoo.com>

<There is nothing magical with the (morpho-)species level here. Couple of years
ago, SciAm printed a cladogram with _H. erectus_ and _H. sapiens_ as sister
species: very fine, except putting them in as unitary taxa assumes a priori
that the former isn't paraphyletic with regard to the later.>

That's a genuine problem, but it hardly ever, if at all, occurs with Mesozoic terrestrial vertebrates.


While it is likely *H. erectus* (as likely with *H. ergaster*) is
paraphyletic with regards to *H. sapiens*, this doesn't affect what David M.
was talking about, in response to what Dave P. was talking about, which was
individual/specimen verus collective OTU capsulization (i.e., putting a
[hopefully] monophyletic group of taxa into a single operating taxonomic unit).

I was actually trying to make Mike Keesey's point, which is that using populations would be ideal. The closest thing we have to populations of Mesozoic terrestrial vertebrates is most likely something like the currently recognized so-called species.


In these, the chosen OTUs are not tending to be inclusive, but subsidiary.

The subsidiarity principle of the EU is that every decision should be made at the appropriate geographic level, so I think you are saying the chosen OTUs are nested within each other. But what do you mean by "inclusive" then?


Closest sister taxa in cladistics are in fact given as the closest known form
to the basal taxa that _could_ have given rise to the more derived sister
taxon, so that, say, *Scleromochlus* represents a descendant of the
paraphyletic group from which also sprang Pterosauria, based on Benton's
phylogeny.

(Which, importantly, is the same as saying that Pterosauria represents ( = is) a descendant of the paraphyletic group from which *Scleromochlus* is descended.)


David M. mentions in the post that he is testing on this method

Not rigorously. I only mentioned one case where I've seen that it doesn't matter much. In general, I do recommend against using large taxa as OTUs, especially if their internal phylogeny is unknown or poorly supported.


I would argue it should be done case by case, but for the most part, exclude
all juveniles and suspect juveniles, and run trees with them with constricted
outgroups to test for variability in the tree.

Or exclude all ontogeny-related characters, but that requires identifying all of them, which is difficult.


They also tend to cause
long-branch attraction, something tested for in living species recently, though
I should not be mentioning this without providing a cite (I am not on my own
computer where I keep my reference works, and I am very tired at the moment).

You are probably thinking of the Systematic Biology paper by Wiens et al. (2005) the title of which begins with "Ontogeny Discombobulates Phylogeny". It shows that the neotenic salamanders tend to form a spurious clade in morphological but not molecular analyses, and that this artifact disappears if they are coded as unknown for ontogeny-related characters.