Fw: Combined answer 1: cladistics

["David Marjanovic" <david.marjanovic@gmx.at>]

Wrote this before I saw that the original had gone to the list, too.

> > Have you seen those studies (for example, Atchely & Fitch 1991, Hillis
et
> > al, 1992) where trees generated via cladistics were matched up with
known
> > phylogenies?
>
> No; are the known phylogenies those of viruses?
>
> > Of the 135,135 possible trees in Hillis et al, all five methods under
> > consideration (parsimony, Fitch-Margoliash, Cavalli-Sforza,
> > neighbor-joining, unweighted pair-group) predicted the true phylogeny.
> >
> > Seems pretty darn accurate in those cases, at least!
>
> Seems too good to be true. Neighbor-joining is phenetics. I must confess I
> don't know Fitch-Margoliash, Cavalli-Sforza or unweighted pair-group...
the
> latter sounds phenetic :-) ... do you have the complete refs?