dispersal fantasy

[pierre deleporte]

Hi, comments in the text below

A 22:17 24/06/2002 -0400, John Grehan wrote :
>I will present some comments here on a paper that was recently drawn to my
>attention and give, in my opinion, and example of an absolutely abysmal
>scientific quality for a 'biogeographic' study. I say this knowing that
>99.9% of systematists practice exactly the same kind of biogeography and
>therefore would accept the scientific credentials of this paper as being
>first class. I acknowledge, therefore, the likelihood that most on this
>list will share that view. However, I make my comments for the few that may
>be seriously trying to think about biogeography in a different way. For
>those who believe my criticisms are misguided, I will be interested in any
>rebuttal.
>
>The paper is by Raxworthy, C. J. et al. 2002. Chameleon radiation by
>oceanic dispersal. nature 415: 784-787.
>
>In the conclusion of the paper they claim that their analyses provide
>evidence for considerable oceanic dispersal of chameleons, and support the
>hypothesis that dispersal, rather than continental break-up, was the
>precursor for species radiation. Far from providing 'evidence' my
>contention is that this was the view the authors read into their
>non-biogeographic data. This 'data' comprised preconceived ideas about the
>historical significance of matching biological area cladograms to
>geological narratives.

All "data" are loaded with preconceived evidence. Call it theory,
background knowledge, or what you like. There are no "data" standing by
themselves. Data are interpretations. Even defining and delineating "data"
is theory-loaded (or meaningless).
The more, interpreting data is (and must be) theory loaded (or is meaningless).

>  According to the authors a (vicariance) geological
>history is supported where biological area cladograms match a postulated
>geological history, whereas the absence of a match is 'evidence' of
>dispersal.

Why not ? This is quite arguable and debatable. Why should'nt there be
convincing geological evidence ?

>  This approach demonstrates a complete absence of any integration
>between biology and geology.

Certainly not. Relying on possibly strong geological evidence for
explanation is in fact a way for such an integration. Let's rather discuss
the arguments.

>  Instead biogeography is simply the narrative
>invented according to a selected geological narrative.

All historical reconstruction is just a tentative explanation, "narrative"
some way. It may be more or less well argued (see also recent de Jong's
message). Arbitrary narrating is flawed, not tentative narration in itself.

>Of course the authors use the propaganda tool of asserting the geological
>narrative is "well documented", but its still a story.

There may be a worse thing that an awkwardly argued story: e.g., a formal
analysis of data without biological or geological justification, i.e.
without explicit reference to background knowledge about evolutionary
processes.
And the propaganda tool of claiming just-so "integration between biology
and geology" without arguing which way it should be done and why is not
worth much consideration.

>  They also appeal to
>dispersal by accepting the precedence of their 'molecular' clock
>reconstructions that are also incongruent with the postulated geological
>ages of the areas occupied by the lizards.

Both are arguable and debatable.

>  They do not bother to consider
>that they may have novel evidence that the geological age is wrong. Neither
>do they consider the possibility that the molecular evidence may be wrong,
>and the lack of congruence between biological area cladograms may be due to
>differentiation independent of the postulated geological splitting events.

Or not.

>Another argument about dispersal also comes from geology, rather than
>biogeography. They claim corroborating oceanic dispersal' by the presence
>of lizards on the Comoros islands that 'never had contact with other
>landmasses leaving only oceanic dispersal (note the similarity to the
>Galapagos dispersal argument. Their contention may in actual fact be
>correct, but notice its based on the acceptance of a geological theory.

Why not ?

>  It is not 'evidence' from any biogeographic analysis.

Which kind, and why is it better in the present or general case ?

>The claim (abstract) that the study 'further highlights the importance of
>oceanic dispersal as a potential precursor for speciation' which may be a
>misrepresentation of their actual work. Far from highlighting the
>importance of oceanic dispersal, their study seems to highlight the
>continued acceptability and credibility of dispersal fantasies based on
>geological and genetic narratives - in a supposedly leading scientific
>journal.

Dispersal may be the real process and not fantasy, rejecting alternative
vicariance hypotheses may be a nice way to argue for dispersal, geological
and genetic narratives may be well supported by background knowledge and
evolutionary theories, and relying on geological or genetic evidence is
arguable and not flawed in itself.

>Food for thought and rebuttal.

Indeed.

best, Pierre