Re: Can we ever recover ancestors? (Was: Fwd: Are dinosaurs really reptiles? (2))

[John Conway <john.a.conway@gmail.com>]

David Marjanovic wrote:
> Mike Taylor wrote:
> > The example that springs to mind is the stratigraphic sequence of
> > centrosaurines in Horner's book _Dinosaur Lives_.  I'm not saying that
> > it's a slam-dunk, but I would at least be slow to dismiss the
> > possibility that they really do constitute an ancestor-descendent
> > chain.
>
> The possibility certainly deserves consideration -- but only if none
> of these centrosaurines (except the last) have autapomorphies. This
> part was routinely overlooked before the introduction of cladistics.

Maybe it was overlooked, but it's also wrong. If you have a matrix of 
300+ characters, surely some will revert in an ancestor-descendant 
chain. Insisting on total lack of autapomorphies will lead you to a 
false phylogenetic result.

The problem at it's core seems to be trying to compare incomparable (or 
vary difficult to compare) evidence: "how complete is this part of the 
fossil record?" versus "how likely is it that this trait reverted?"

Apples - oranges = ??? (and don't say "ppl", that'd be lame.)

--
Palaeontography: http://palaeo.jconway.co.uk