[Taxacom] Drosophila melanogaster name change?

Kim van der Linde kim at kimvdlinde.com
Tue Apr 13 09:28:41 CDT 2010


Hi Richard,



On 4/10/2010 5:19 PM, Richard Zander wrote:
> Regarding the paraphyly of Drosophila, the diagrams at the Web site
> certainly show (in the caulistic sense) that Drosophila is a
> progenitor (as one or more species diagnosable as Drosophila s.str.)
>  of several autophyletic genera. Okay, that is what it shows.

Yeah, phylogenetic trees are just that, a diagram of relatedness...

> My take on molecular analysis is that it shows (ignoring
> reticulation and other problems) genetic continuity and isolation of
> lineages, but not speciation events because ancestors may survive
> speciation events.

Duh, separation of lineages is of course not the same as speciation, and
most of the times, at least one of the ancestors will survive the event.

> They do, they do, as is abundantly demonstrated by the myriad
> homoplastic species, genera and families making phylogeneticists so
> uncomfortable they split and lump to hide what are commonly actually
> multiple molecular lineages of the same taxon in phenetic stasis.

And here we arrive at the question, what characteristics do we use to
determine whether two clades are a single taxon or multiple taxa. For
most of history, we assumed that phenotypic differences mirrored genetic
differences, but we realize more and more that this is not necessarily
the case. So, we end up with the situation that we have to decide how to
determine what are different taxa and what are not.

> So... If your autophyletic genera are robust in expressed traits
> important in evolution, then you have a good new classification, and
>  the model fruit fly should be renamed with a correct name in
> Sophophora.

Agreed.

> One can of course use the nomenclatural synonym "Drosophila
> melanogaster" as a kind of taxonomic DOI.

Yup, and this is also the easy way to deal with databases etc. Large
databases such as google already search automatically for other search
terms that are covering the same thing, for example spelling variants of
American and British, etc.

> What does "robust" mean? If the lineages chosen as molecular
> autophyletic genera are simply largely random sets of species, then
> one can usually find a combination of traits that by coincidence
> serve as synapomorphies for a group. This does not diagnose a robust
>  taxon.

Agreed.

> If the trait combinations (sure, polythetic is okay) are
> significantly better for the recognized genera than any other groups
>  formed by random combinations of species, then you have a robust
> genus that avoids the multiple comparisons problem in statistics.

What we have now is a group of species that is primarily determined by 
the history of the genus, without any sensible criteria for its 
circumscription. In fact, there is not a single synapomorphy that holds 
the genus together, primarily because of the huge variation between the 
species. Over time, various clades have been lifted out the the genus 
based on some unique morphological character, but lacking the robustness 
that you talk about. At the same time, obvious clades that are really 
different and robust in many ways are left in the genus not based on 
scientific arguments, but primarily based on personal preferences, such 
as the Hawaiian drosophilids.

> If you want to present your research as "hard science" then you need
> to quantify the statistical analysis and do a Bonferroni
> calculation, or just a joint probability will do since this stuff is
> necessarily Bayesian as a one-time historical reconstruction.

Taxonomy and 'hard science', hummm.....

> If this seems complex, of course it is.

Nah.

> The problem with classification nowadays is the extreme simplicity
> of the evolutionary analyses on which a phylogenetic classification
> is based. Problematically, a phylogenetic classification does not
> help professional evolutionists do research in evolution because the
>  simplistic assumptions are built-in and affect the results of
> research.

Well, phylogenetics is a tool to do taxonomy, it is not the end all 
answer to everything.

Kim




More information about the Taxacom mailing list