Australopithecus and orangutans

[John Grehan]

At 09:36 AM 5/15/2003 -0700, Julian Humphries wrote:

>>My understanding and practice in morphology is to designate apomorphic
>>characters.
>
>That is simply wrong.  The majority of workers (morphological or
>otherwise) do not a priori designate a character state (not character),
>as apomorphic.  How could you?

If this is true I may have somehow missed the boat on what people in
general are doing. For the orangutan question if I am looking at the
interrelationships of humans and apes and took a very strict approach I
would not include characters that were also present in the outgroup (the
rest of primates, even mammals in general). Thus through outgroup
comparison I am in a position to propose apomorphies for analysis before
the analysis itself is carried out (to determine the 'best' grouping of
those characters). Even though by definition the rejected group of proposed
apomorphies are now designated as homoplasies and may be considered
parallelisms, convergences, plesiomorphies etc., that designation is only
by default of the analysis in question. They still remain potential
apomorphies and may well be corroborated in future studies (I saw this
happen once with a lepidopteran group were adults were used and a phylogeny
proposed, but once the larval stage was discovered the authors found
themselves forced to change their classification entirely and relegate
their former synapomorphies to plesiomorphies etc).

I would not use characters to analyze great ape relationships that are
present in the outgroup. I would designate these as potential
plesiomorphies and leave them out of the analysis. Otherwise I would not be
doing a cladistic phylogenetics as I understand it.

With respect to being strict, it is possible that one might be less strict
and use a character that broke the exclusion from the outgroup if one was
willing to argue independent origin (e.g. treat thick dental enamel as a
shared derived character for humans and chimps even though it occurs in one
Old World and one New World monkey species).

>There is absolutely no fundamental difference between
>molecules and morphology as to the assessment of "shared derived
>characters."

Only in that there appears to be no exclusion of characters that are not
confined to the ingroup for which relationships are being analyzed.

>In fact, the situation is made more complex due to optimization of
>character state evolution (equally parsimonious solutions, even on the
>same tree, may yield multiple distinct sets of shared derived
>characters, thus synapomorphies may depend on which optimization
>strategy you adopt.)  Thus, a state appearing in an outgroup and an
>ingroup may be a retained pleisiomorphy, or a novelty, depending on
>optimization. Characters with homoplasy, may, at some level, still
>diagnose major clades.  All of this complexity is why that "computer
>contrived analysis" (this is the kind of word baiting typical of
>Kinman), is so necessary.

Baiting or not (and it was not my intention to use the term in that
context) I was referring to the term 'contrived' as a descriptor for the
very situation described above since the outcome can be so much a matter of
how one might select and apply different algorithms along with the
assumptions themselves. I will use the term 'contrived' even though I may
also use such methods.

Well, I admit my lack of expertise on phylogenetic analysis and perhaps I
am making a fool of myself in my ignorance. I will have to live with that,
and perhaps learn along the way.

John Grehan

Dr. John Grehan
Director of Science and Collections
Buffalo Museum of Science
1020 Humboldt Parkway
Buffalo, New York 14211-1293
Voice 716-896-5200 x372
Fax 716-897-6723
jgrehan at sciencebuff.org
http://www.sciencebuff.org/biogeography/Panbiogeography/Panbiogeography-Gate.htm
http://www.sciencebuff.org/HepialidaeGate.htm