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

A comment on cladistic characters

        Being a HUGE fan of vertebrae, I am always disappointed at their
treatment in cladistic character lists.  Showing their bias against vertebrae,
sauropod character creators often score each metacarpal, metatarsal and even
(on occassion) phalax for the presence or absence of a character (say a
tabular process on the lateral border) based on positions (I, II, III, IV, V)
BUT then they list a character like "tall neural spines in proximal caudal
vertebrae" or "dorsal vertebral supradiapophysial laminae meets
suprapostzygapophysial laminae at midline 2/3 way up neural spine" and this
character is then magically believed to apply to all of the vertebrae in the
"region" being referred to, be it the undefined proximal caudal region or the
dorsal column, etc.  The attribution of characters to the vertebral column
must be done with the utmost of care.  A recent paper speaks of a
hyposphene/hypantrum articulation being present in one taxa but not another,
yet putting my hands on both taxa revealed that the hyposphene/hypantrum
articulation in these two animals is based purely on position, much like the
presence or absence of a limb character.
        This mistreatment of vertebrae in sauropod cladograms distresses me to
no end, as sauropod vertebrae possess an extremely large amount of
phylogenetic signal within their axial skeleton, it just takes an almost
inordinate amount of time to map each vertebra's laminae and determine which
characters are variable and which are extremely useful.  Sauropod vertebrae
can be identified to an "exact" position, with dorsal vertebrae being
confidently identified to within 1 position of reality and caudal vertebrae
within 2 (I would say with relative ease but I know how much time I have spent
learning this neat trick).  I feel that if limb material gets a "MC I,
character x is present/absent; MC II, character x is present/absent" then
vertebrae should be afforded the same rights with "Position 1, character x is
present/absent; Position 2, character x is present/absent...".
        I will be the first to add that this is impossible with taxa where
there are not complete specimens to compare against, so maybe the above is
somewhat off-base, however much of what I have been reading does not actually
correspond with reality of the material at hand.  There are certain "rules"
that apply to vertebrae that carry across all sauropod taxa (such as location
of parapophysis to determine position) that are either unknown or ignored by
researchers, which leads to erroneous literature.
        A cladogram should be able to withstand 1 or 2 characters that are
simply wrong, but what happens when the majority of vertebral characters are
suspect?  And what value are cladogram characters that are functionally
driven?  How would PAUP react when of 68 characters, 15 are vagure
generalities regarding vertebrae that are often false elsewhere in the
vertebral region being specified _and_ 12 more are clearly functional or
sexually dimorphic in nature (like the degree of neural spine recurvature in
caudals 10-12 of Diplodocus).
        I agree with Nick Longrich that quite often researchers today get
caught up in the battle frenzy that a cladogram can generate, for here is a
way to take data lifted right out of papers and pictures, make some decisions
(hopefully wise and well thought out) and, wahla, a bona-fide scientific
phylogeny fit for publication.  This is a seductive power.  Have a low CI?
Add more characters!  That'll fix it!  And so on.  The fear is the researcher
soon loses sight of the character selection and starts scoring odd, ill (or
un) defined characters in order to generate more characters because the more
characters they have the more accurate their cladogram must be.  The cladistic
method has led to a variety of new papers on sauropod phylogeny, leaving
careful descriptions far behind.  Yet it is the alpha taxonomy that needs to
be very sound for the cladograms to work!  Using the Ultrasauros holotype
dorsal vertebra in a cladogram should have sent off warning signals that
proclaimed, "uhm, this thing is falling out in the diplodocid section, maybe
there is a problem?" yet this did not happen.  Which concerns me because I do
not exactly know why this is the case.  Either folks did not run the beast
through their software, rather inserting it afterwords on their cladogram with
the thought "thats where it belongs" or their characters were so basic that
they completely "missed the boat" with vertebral character selection.  Each of
which is pretty poor science.  I know the field is "research", to search
again, and that is what makes this job so fun.  Yet when research becomes
"rehash" then troubles abound.
        Brian Curtice