More on Biogeographic memory

Pierre Deleporte Pierre.Deleporte at UNIV-RENNES1.FR
Tue Oct 9 19:50:37 CDT 2001

Dear all arguing in and following this thread,

It happens that I recently had to present historical biogeography methods,
and I decided to qualify most of Panbiogeography (CGH 1999 version) as a
"descriptive" approach. The term may be awkward, but I simply mean that
connecting by shorter straight lines the localities harbouring the same
taxon looks at first sight as a mere alternative to drawing the contours of
the area containing these localities. This procedure does not involve in
itself any historically meaningful dimension, and neither does in itself
the identification of some redundant pattern among these lines or "tracks",
nevertheless this could constitute the first step of an approach using
biogeographical "spatial distribution networks" some way.
But why should biogeographical "units" be defined this way rather than any
other way, this I could not find in CGH 1999. I suggest that this reveals
some lack of clarification of the paradigm (see below).

I also contrasted historical biogeography with phylogeny reconstruction,
and the problem of elementary areas and taxa delineation (in biogeography)
with that of terminal taxa and characters delineation (in phylogenetic
I understand quite well the insistence of John Grehan that a biogeographic
unit (area or else...) should better not be arbitrarily defined a priori,
but rather emerge from some explicit analysis. Rosen (1988) already
insisted on this point about PAE (in this case, hierarchical grouping of
unit localities comes from overall similarity between the biotas in these
localities in terms of presence / absence of species); Hovenkamp (1997)
also insisted on avoiding arbitrariness in defining "areas of endemicity"
in vicariance biogeography, and his method avoids forcing biotas into such
predefined areas.

The latter point seems to rejoin some way Morrone's allusion to the
"primary homology" problem in phylogenetic systematics (defining and
delineating what is a priori a "similar character" in different taxa: the
"bete noire of phylogenetic inference" according to Pogue and Mickevich
1990), but in my view this could also be contrasted with the problem of
delineating terminal taxa in systematics (whose monophyly should be a
priori asserted, just like the "biogeographical consistence" of areas). The
general problem is to avoid arbitrariness and irrelevance in the
delineation of elementary units used in the analysis. But of course there
is no "theory neutral" delineation (or "observation"), and alternative ways
to define and delineate elementary units should be evaluated for their
consistence with the logics of a given paradigm... provided that this
paradigm (theories and background knowledge, goals and methods...) has been
stated clearly (this vails for areas of endemicity, distribution ranges,
spatial minimum spanning networks, baselines, tracks, main massings and
what I know, as well as for terminal taxa and characters - character states).

Now it seems clear to me that an elementary notion of phylogeny was
introduced in CGH 1999 Panbiogeography : the suggestion of drawing the
shortest straight line between the distributions of two sister groups (thus
not only inside one taxon). But Hovenkamp is correct in stating that the
book does not explain how to implement this when dealing with the
"hierarchy of sister-groups" constituted by a phylogeny, and its relative
temporal dimension (which Hovenkamp 1997 attempts to take explicitly into
account in his own approach).

Nevertheless I was striken by the fact that when Craw et al. draw the
shortest straight line between the distribution of two sister groups,
Hovenkamp draws a separation between adjacent distribution areas of two
sister groups. The line or track has to cross the separation or barrier. On
this limited point, the two approaches look very similar, and this could
give a handle to contrast them (Peter : why do you draw your line between
the spatially closest borders of the extension ranges? Or are there
exceptions to this rule?).

Finally, I can't resist quoting Hovenkamp (with permission?...) and suggest
that we will be really on the way to finding methods in biogeography when
we will be able to state properly what is(are) the question(s) we ask
(hence my allusion to the "paradigm" above).

I'm convinced that the debates in phylogenetics can help clarifying
this(these) other historical approach(es): biogeography(ies). In both cases
we want to reconstruct the past of an evolving system, and we have some
knowledge of the end of the story... thus we have to implement some
explicit "transformation rule" for retrodiction, and justify the whole
stuff a logically consistent way, including relevant rules for delineating
elementary items.

We are not out of the wood yet.

thanks for this stimulating thread,
best and cheers,

At 09:38 09/10/2001 -0500, you wrote:
>Dear Curtis (an all),
>as I conceive it (although panbiogeographers as well as cladistic
>biogeographers will disagree) main differences between panbiogeography
>and cladistic biogeography are not only methodological, but their practice
>seeks to answer different questions.
>I have began to explore these issues in an article with J. Crisci
>(1995, in Annual Review of Ecology and Systematics), and posteriorly
>elaborated it further. An article in Diversity and Distributions, which
>will appear very soon, will present this in terms of primary geographic
>homology (panbiogeography) and secondary biogeographic homology (cladistic
>With best regards
>Juan J. Morrone
>Fac. Ciencias-UNAM
>Mexico city

More information about the Taxacom mailing list