[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: Thoughts on the biggest morphological bird analysis

David Marjanovic wrote:

Of course, but when different genes say the same thing, without an identifiable selection pressure that could work on them all (like base composition bias), I'll trust large molecular analyses over fossil-less or otherwise too small morphological analyses any day of the week.

The thing is, the source of the homoplasy may not be immediately identifiable. There may be (and often are) more subtle factors at work than simple compositional bias. Compositional bias, if present at the DNA level, might not be a product of any selection pressure (at least not for vertebrates). Other biases work at the amino acid level, for example, and are more insidious - and so they tend to be more difficult to identify. Remember, the analysis is only looking for non-randomness in the dataset. It has no way of differentiating the 'true' (historical) phylogenetic signal from any biases (homoplasy) that are also present.

With morphological analyses, on the other hand, the putative homoplasy is often plain to see. Like the grebe-loon clade, where foot-propelled diviing is the prime suspect. Or the bat-cologo clade, where it's aerial locomotion.

Thus, a bat-colugo clade is clearly out of the question, and a loon-grebe clade looks unlikely -- here we have only three molecular analyses so far, but we have Mayr's (admittedly small) morphological analysis that includes the extinct flamingo diversity and *Juncitarsus*, while Livezey & Zusi only use extant taxa in that part of the tree.

I agree with you. To some degree I'm playing devil's advocate here, because many heavy-duty molecular analyses aren't as "case closed" as they're made out to be. However, you can't rule out the possibility that a morphological analysis with an expanded data set (i.e., more fossil taxa) might actually confirm these clades, such as grebes and loons grouping together. As you say, these hypotheses require further testing.



Check out all that glitters with the MSN Entertainment Guide to the Academy Awards® http://movies.msn.com/movies/oscars2007/?icid=ncoscartagline2