Re: BCF (was New Article in Experimental Zoology)

["David Marjanovic" <david.marjanovic@gmx.at>]

> Yes, Ax makes for some fun reading. Dunno why his results are any more
> suspect than computerized cladistics, though.

Because he uses very few characters. Because it can often be argued that two
characters he lists are the same respectively interdependent. And because he
doesn't know that his big picture is the most parsimonious tree that can be
grown from his own data. He doesn't make a data matrix; he takes one taxon
and looks for its sister group, and then he looks for the sister group of
the first two together. This goes so far that he states that the sister
groups (sorry, adelphotaxa) of Animalia and Tardigrada cannot be found (not
just that he can't find them). Of course this search is somewhat
scenario-based. This way he finds Articulata, composed of Annelida and
Arthropoda, and Arthropoda instantly reverses several articulate
synapomorphies... I haven't tried, but maybe Ecdysozoa and Lophotrochozoa
are more parsimonious even when only his data are used. (It must be said
that he really knows the animals he writes about. For example, he does find
Platyhelminthes and Nemertini as sister groups, because of IMHO suspect
characters, but not because of any of the traditional characters, such as
the presence of parenchyma, which are not present in the basal
representatives. And he finds Mollusca and Kamptozoa as sister groups
because kamptozoan larvae are very mollusc-like.)

> There is, I suppose, as little basis for using
> his particular character-weighting scheme as there is for a
> weighting scheme that is a no-weighting scheme.

I'm not under the impression he really weights... well, he does, but only as
1 or 0, he dismisses e. g. all molecular data and all morphological
characters he thinks are not reliable... anyways, not weighting at all
standardizes the procedure, it makes different cladograms more comparable.