<< And to add a dinosaur thread, the new find also shows a mosaic of traits seen in later hominids, evidence that the traditional linear model is breaking down for human evolution just as it seems to break down for the evolution of birds into dinosaurs. >>
The so-called "linear model" is and always has been nothing more than an artifact of arranging taxa in a cladogram. To convert any cladogram into a "linear model," or Hennigian comb as it's sometimes called, simply single out one lineage to be the spine of the comb and arrange all the other lineages as "side branches" of the selected lineage. Singling out the lineage that leads to man, for example, produced the notorious "scala naturae" that dominated phylogeny for circa 200 years. So the model is not breaking down; we're simply finding lots more side branches. As for mosaicism, if you get enough of it, it makes practically all cladistic analyses suspect. Too much convergence turns the cladograms into mush.