Re: Origin of Birds

[Bill Adlam <wa105@mead.anglia.ac.uk>]

Ralph Miller wrote:

>As I understand it, cladistics are based on the synapomorphies (shared
>derived traits) between animals, inherited from common ancestors.  Cladists
>don't necessarily concern themselves with precisely who begat whom.

That's kind of right and kind of wrong.  There are two types of cladistic 
taxonomy, the ordinary kind which most taxonomists simply call 'cladistics', 
and a variant with the same methods in practice but a slightly different 
philosophy usually known as 'pattern cladistics' or 'transformed cladistics', 
but its practitioners simply call it 'cladistics'.  Clear?
 
Ordinary cladists are very much concerned on precisely who begat whom.  Their 
aim is to reconstruct phylogeny - the evolutionary sequence of diverging 
lineages.  Synapomorphies are used to group pairs (occasionally larger sets) 
of species together, on the grounds that they both descend from a particular 
ancestor.  The synapomorphies for the group can be used to reconstruct some 
characteristics of the 'hypothetical common ancestor'.  This is the organism 
which is inferred from the analysis to have been the ancestor - the 
'hypothetical' indicates that it is what is predicted from the analysis, 
rather than (necessarily) a known species.  

Once you've got an HCA for two species, you can group them together, replace 
them with the HCA and go on to calculate what it's relationships are.  This 
is all done automatically by computers.  However, cladists don't ask 
palaeontologists to go out looking specifically for the ancestors they've 
predicted.  The reason for this is the sparseness of the fossil record: most 
species that have existed are unknown to science and probably always will be.  
The predicted ancestor may well really have lived, but it might have been 
restricted to a small area, or to a short time, or might have been too rare to 
have much chance of preservation.

Cladists don't seem to mention HCAs very much now, perhaps because they are 
unlikely to be found.  And even if one species is descended from another 
known species, this is almost impossible to prove, so the hypothesis is very 
hard to test.  It takes a good fossil record and a lot of very careful study 
before we can be sure that one species descends from another.  It's often 
said that one hominid species is an ancestor of another, but in most cases 
there is a good chance that another unknown species existed which is the real 
ancestor.

The older literature often suggests that one species is ancestral to another -
this would be frowned on nowadays by cladists and, I think, by evolutionary 
taxonomists as both unlikely and very hard to test.  Evolutionary taxonomists, 
who allow paraphyletic taxa, have equal difficulty identifying ancestral 
species.  However, they can often specify a paraphyletic group, and state that 
an organism had ancestors in this group.  Cladists can say something very 
similar by saying one clade is a subclade of another.  

Now pattern cladists.  Like ordinary cladists, they base their taxonomy on 
phylogeny, as reconstructed from synapomorphies.  But they use the word 
'phylogeny' in a special sense.  I have never been able to determine what it
is they mean by it.  My taxonomy lecturer was a pattern cladist, and all we
got out of him was that "of course" phylogeny wasn't anything to do with 
evolutionary history.  Can anyone enlighten me?

So, to summarise, ordinary cladists are concerned with ancestors, as are 
evolutionary taxonomists.  Pattern cladists probably aren't.  However, the 
vocabulary of evolutionary taxonomy allows assertions like "dinosaurs are 
descended from thecodonts", whereas the cladistic equivalent would be 
"dinosaurs are archosaurs".

>                                                                     How
>could we determine whether some theropods descended directly from other
>non-avian theropods or were secondarily flightless descendants of flying
>forms?  In other words, how might we determine the likely anatomy or
>identity of the aforementioned common ancestor?  I suppose the jury will be
>out until more fossil evidence comes in, but, in that event, precisely what
>evidence would we need to settle the matter (assuming that we could set
>aside our biases)?

As mentioned above, we can reconstruct the characteristics of the hypothetical 
common ancestor of two theropods A & B.  If it turns out that A has all of the
characteristics of the HCA, it is possible that A is in fact the ancestor of 
B, but this is very speculative.  The fossil record would have to be much, 
much better than it is before we could be sure there wasn't a C which was the 
ancestor of A and B.  If the fossil record was that good then stratigraphy 
would also be very useful here - as it is, it isn't.

As for theropods being secondarily flightless, it would be very convincing 
evidence if a theropod was found with asymmetrical feathers.  Finding them in 
distantly related theropod groups would show that most or all were descended 
from birds.  Such a discovery would be unlikely, though, even if theropods are 
in fact birds.

George Olshevsky will doubtless tell you why he thinks Birds Came First, and 
what future evidence could support his theory.

                                                        All the best,

                                                                        Bill