DIGIT LOSS AND GAIN

[Tetanurae@aol.com]

George Olshevsky wrote:
<<[1] Nobody has found an unassailable instance in dinosaur evolution of the 
reappearance of digital phalanges, or of the reapparence of a metapodial once 
lost. (Also true of mammalian evolution, I believe.) This is actually rather 
odd, because supernumerary vertebrae--particularly cervicals and caudals--do 
appear from time to time in dinosaur lineages; why not supernumerary 
phalanges?>>

I don't believe that this is in fact true George.  In marginocephalians, 
there are many instances of possible metapodial phalangial loss and then 
gain.  In _Heterodontosaurus_, which is probably at the base of the 
Marginocephalia stem, the manual digital formula is 2-3-4-3-2, in 
_Psittacosaurus_, it's 2-3-4-1-0, in _Leptoceratops_ it's 2-3-4-3-1, and in 
_Protoceratops_ + Ceratopidae it's 2-3-4-3-2.

Most phylogenies of Marginocephalia have this basic structure:

-+--Heterodontosaurus
 +--+--Pachycephalosauria
    +--+--Psittacosaurus
       +--+--Leptoceratops
          +--+--Protoceratops
             +--Ceratopidae

And now with manual phalangial counts:

-+--Heterodontosaurus [2-3-4-3-2]
 +--+--Pachycephalosauria [unknown]
    +--+--Psittacosaurus [2-3-4-1-0]
       +--+--Leptoceratops [2-3-4-3-1]
          +--+--Protoceratops [2-3-4-3-2]
             +--Ceratopidae [2-3-4-3-2]

So either Psittacosaurus and Leptoceratops greatly reduced manual digit V 
independantly, or it was lost and then regained once ceratopians started 
walking around quadrupedally all the time.  The second option is actually 
most parsimonious as it requires only one "step" (a supression of the gene 
that told digit V not to develop).

This topic was originally brought up in the discussion of prosauropods being 
sauropod ancestors or sister groups, and I cannot see the difference between 
manual and pedal phalanges in this sort of scenario.

To dismiss the cladograms that state that prosauropods are a paraphyletic 
assembly of progressive sauropod outgroups simply because you claim that 
digits can't re-evolve (with no other proof besides "I've never seen it") is 
at the very least unscientific.  In fact, it's quite a circular argument: 
 I've never seen a case where digits could have re-evolved so it can't 
happen, you're scenario that shows digits re-evolving is impossible because 
digits can't re-evolve.

Looking forward to more discussion,

Pete Buchholz
Tetanurae@aol.com