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

Re: Birds and mosasaurs [Rather long and theoretical]



     I have to admit I had a hard time following the argument you
presented here.  It might be better to let Jonathan Wagner or Tom Holtz
field this one. 

> Cladistic analysis deals very well with the case: A+ B+ C+.  What do we
> conclude for A- B+ C+?  That there is an ancestral form A'+ which we've
> missed and that A is not part of the clade?  That's fair enough.  But what
> about the case here where we have birds (Z+++), and Maniraptorans with odd
> assemblages like B-+-, C+--, D--+ , but A---.  (This grossly oversimplifies,
> but bear with me).  We can declare, for example, B the winner based on
> statistics, but we know that this means C and D independently developed
> characteristics that are found in Z but not in B.  Alternatively, we can
> assert that we've missed A'+++ as before.  Neither is very satisfactory.  

     You have said in other postings that you think bird origens lie
in SOME theropod group, so we'll pretend that this is the case and
just deal with the best bird ancestor candidate among the theropods.  
     First of all, as I've pointed out in other postings, the confusion
in nailing down the theropod group closest to birds dpoesn't set in 
until we have already gotten to the coelurosaurs (coelurosaur and
maniraptorian mean _almost_ exactly the same thing). 
     To put it another way, you called birds "Z+++", and various
maniraptorian groups "B-+-", C+--, D--+, and A---.  The problem with this
simplification is it doesn't take into account the big list of "+"s that
match up before we get to the specific coelurosaur families, so lets put
those in.  Lets call birds:

             Z+++++++++++++++++++++++++++++++++++++++++

      With the last three pluses being the ones we have already
established.  Lets call primitive theropods:

             T+++++++++++++++++------------------------ 

     Pleisiomorphically, they have of bird-like traits, but are missing
quite a few.  Tetanurans would be:

             U+++++++++++++++++++++++++++--------------   

     And lets say group A represents primitve COELUROSAURS:

             A++++++++++++++++++++++++++++++++++++++---

   With two other coeurosaur families (dromeosaurs, troodontids) being:

             B++++++++++++++++++++++++++++++++++++++-+-
             C++++++++++++++++++++++++++++++++++++++--+

     Obviously, I am doing some gross oversimplifying of my own, but the
point is the confusion in pinning down the specific coelurosaur group is
minor compared to the BIG list of relatively unequivovical synapomorphies
that get us to Coelurosauria in the first place.  Group "T" is your "basal
theropods", and despite the confusion presented by the last three of my
hypothetical characters is minor compared to the big list of
synapomorphies that dromeosaurs and troodontids SHARE with birds, that the
basal theropods do NOT have.   

> That is, it may be possible for an
> organism to have a suite of *genetic* characteristics which are "bird-like",
> in the absence of  an avian phenotype -- simply because they can mutate in
> that direction through viable intermediates.  

     I am not familiar with Raaf's work, so I can't comment on it, but 
what you are saying seems a little confusing; are you saying that an
organism that isn't related to OR looking like a bird could mutate so
that it had some GENETIC similarity to a bird, but still didn't LOOK like
a bird?  This wouldn't muddle a phylogenetic analysis at all; if the
convergent traits (in this case, strictly GENOtypic) can't even be
detected in the fossil record, they can't muddle up a phylogenetic
analysis.
     If you are talking about convergent evolution creating superfical
similarities in unrelated species, I think you have it backward.  The
similarities between unrelated species are _phenotypic_, NOT genetic.  And
again, this doesn't become a problem for birds and theropods until we have
already nailed down the coelurosaurs in general.          

> If you buy this, then the A--- species may be closer to Z+++ than any of A's
> descendants, because each of the genotypes giving rise to the "-" phenotype
> can evolve in the direction of "+" through developmentally and ecologically
> viable intermediates.  

      But again, A ISN'T closely related to Z, and if its convergently
evolved genetic similarities aren't even expressed phenotypically
(through skeletal morphology), a paleontologist examining its remains
would have no reason to think that it was.

> Phylogenetic distances can't really be measured directly, because they are a
> complex product of genotypic constraints, developmental constraints, and
> ecological constraints.  

     Phylogenetic trees using skeletal morphology still seem to match up
pretty nicely with those derived by geneticists, at least in the case of
living organisms.  This seems to imply that skeletal morphology (read
"phenotype") IS usually pretty diagnostic of genotype.            

> However, it is just such constraints that make the
> idea of phylogenetic space useful, albeit at the risk of reintroducing
> concepts like "evolutionary trajectory".  Minus any teleological component,
> such terms simply describe a case in which genetic, developmental and
> ecological constraints together permit evolution in one direction more
> easily than in another, i.e. proximity in phylogenetic space.

     Okay; if you mean "certain designs for doing certain things are
better then others" you are talking about CONVERGENCE.  If you are looking
at CLOSELY related organisms (like the different theropod groups) that
inherited a common starting body design, and this body design is
most easily modified for doing certian sorts of things in only a few sorts
of ways, then you are likely to see different closely related groups
develop thier common body plan in similar kinds of ways.  For this we
use the word PARALLELISM.    

> The catch is this.  Since we're talking about hypothetical ancestor species
> which are close to birds in phylogenetic space, but not in phenotype, we are
> left somewhat uncertain about how far back that genotype existed.  We can't
> test directly for genotype, and might not recognize the similarity (or
> rather the potential similarity) even if we could. 
> Its pretty hypothetical, but I'm > attracted to the idea because it
> explains the strange, mosaic evolution of avian characteristics in the
> coelosaurian radiation AND explains why we see
> these characteristics popping up anew, in different combinations, and in
> different lineages.
  
     I think I finally understand your reasoning; you are saying that the
genes that would code for the avian phenotype that we are calling
bird-COELUROSAUR synapomorphies actually popped up in primitive theropods
but were not expressed at first.  They later BECAME expressed
phenotypically independantly in birds and coelurosaurs, right?  
     This seems really pretty strange to me; I have heared of genes that
USED to code for phenotypic traits being INactivated, but I have never
heard of a random mutation starting out as an intron and then somehow 
being activated later to cause perfectly functional phenotypic trait
without natural selection having ever been involved.  A bunch of genes
just happen mutate in a way that would make a perfectly functioning
phenotype, and then just sit there until the right environmental
conditions arise to activate them?  It sounds like you have taken natural
selection out of the equasion entirely; the genes just KNOW what will work
someday without having ever been subjected to selective pressure; how
could they, if they were never expressed phenotypically when they first
appeared?              

LN Jeff