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


George Olshevsky wrote:
>You have no basis for saying this at all, since there is really no way to
>confirm that any cladogram is true.
        However, if you have a large number of cladograms which show
reversals, it seems likely that reversals do occur.

>What, for that matter, qualifies as a "reversal"? A derived state that
>resembles a primitive state? Or the genuine reappearance of the primitive
>state? And how would you tell?
        All a reversal is (for the purposes used here), is an instance where
a character which has transformed lower on the tree transforms back to the
earlier state at a higher node. The upshot of this is that "reversal" is a
classification of biological effects seen on a cladogram, and need not be
associated with one particular biologic process or another.
        Evolution *doesn't care* whether it is ressucitating old genetic
code or breaking new ground when an adaptation comes about. The evolutionary
event (or group of linked events) simply occurs. It may be easier one way or
another, but how do you tell? I can certainly say that your first option can
not be any less difficult than a "convergence", since that is what it is:
merely a convergence with an ancestral form.
        We observe character states in taxa and classify them as such. We
must test the hypothesis of homology of these character states. We cannot a
priori state that one or another state is homologous, nor can we speculate
upon the processes which lead to the evolution of the character in question
until homology has been tested. We test homology using a phylogenetic
hypothesis (a cladogram).
        If the quadrudactyly of therizinosaurs is inherited from an ancestor
hypothesized to be quadrudactyl, it is homologous with the quadrudactyl pes
of prosauropods. If it is inherited with modification (evolved) from a
tridactyl theropod ancestor, the occurence of four toes in therizinosaurs
and prosauropods is homoplastic. Since the character state was undoubtably
"four toed" in an ancestor of the therizinosaurs (and indeed all dinosaurs),
then the homoplaisy is termed a reversal.
        This reversal may have come about by scraping together the old
genetic code for quadrudactyly, or by the independant novel introduction of
four-toed feet within the clade. Regardless, it is, in terms of the
characters at hand, a reversal, because the character has gone back to its
original state. In either case, under a phylogenetic definition of homology,
the character state itself is not homologous with the original quadrudactyl
state of dinosaurs. Whether it is or is not a true genetic reversal does not
matter in our coding and treatment of the character and character states

        I have no idea how you would tell the difference between the two
options you give. It may only be possible to tell by actually observing the
evolutionary event. The two situations may be indistinguishable in some
cases. One may happen more frequently, or not at all. I doubt this, both
avenues seem likely.
        For instance, I can say that the reversal of the character state
"four toed foot" in therizinosaurs appears to be a reversal of the second
kind, in that the metatarsal has secondarily reelongated to partially reach
the ankle as it did in days of yore. However, the functional aspect of
qaudrudactyly in therizinosaurs is also accomplished by the shortening of
the pes. This may be reappearance of a primitive state, but seems likely to
be a new state. So "four-toedness" in therizinosaurs could be regarded as
both. But note I said "seems". How would I prove either case?

>Take convergence over reversal every time as a parsimonious working hypothesis.
>Or a brand-new lineage.
        This is a philosophical matter concerning character optimization. If
you optimize to avoid reversals, fine. However, if you then reject trees
showing any reversals, you are unduly biasing your result based on a notion
of one type of homoplaisy being "better" than another (see below). You are
going even further than this, you just outright refuse to accept evidence
for the theropodan nature of therizinosaurs, based on your refusal to accept
one reversal. This is clearly overreacting.
        Further, as I note above, since one of the possible avenues of
reversal which *you* list might be viewed as a type of "convergence" rather
than a "genetic reversal" (as I term your other option), how can you say
that all reversals are less likely than convergence?

>Reversal should be considered a last resort, an extraordinary claim
justified by extraordinary proof, not business as usual.
        As opposed to convergence, which is by no means extraordinary? Where
do you get your data? What is your evidence that reversal is somehow
difficult? If you look at each evolutionary change as a phenomenon, what is
it about reversal which makes it less likley? Is it harder to cover old
ground than to break new ground? Is it less likely that I look like my
grandfather than I look like you?

>As I have repeated here until blue in the face: bullatosaurs have an inflated
>parasphenoid, _Erlikosaurus_ has a pneumaticized basisphenoid. These are
>completely different skull bones [...]
        Which (if true) proves nothing. What mattered to the animal was the
pneumaticization, not necessarily which bone it was in. Do you consider the
antorbital fenestra of some archosaurs non-homologus with that of others
because in some it is bordered dorsally by the maxilla and in others by the
premaxilla? Do you consider the orbit to be a different feature when is
bordered anteriorly by the lacrimal or the prefrontal? Characters are
(generally) functional adaptations, and as such the bone can "flow" around a
particular feature, be it a bulge, bump, foramen, etc. as long as the
function is preserved.
        The only way to prove or disprove the homology of the Parasphenoid/
basisphenoid inflation is by testing character distributions within a
phylogenetic hypothesis. A priori declaration of non-homology is not convincing.

>Because if this isn't convergence, and segnosaurs are indeed maniraptorans,
>then maniraptorans must have evolved before the beginning of the Jurassic
>(that's what the jaw in the July 16 _Nature_ says)
        I am not quite clear on the stratigraphy here. Does the article say
*lowermost* Jurassic? Because if it doesn't, I do not see how this follows.
A therizinosaur in the Lower Jurassic says that maniraptorans (as a
potential sister-group to therizinosaurs) must have evolved by the time
those beds were deposited. It says nothing about *before* the Jurassic.
Several million years can make a big difference in such assumptions.
    Jonathan R. Wagner, Dept. of Geosciences, TTU, Lubbock, TX 79409-1053
                    "...To fight legends." - Kosh Naranek