More on Biogeographic memory

P.Hovenkamp Hovenkamp at NHN.LEIDENUNIV.NL
Thu Oct 11 14:16:04 CDT 2001

At 12:26 PM 10/9/01 -0400, John Grehan wrote:
>(...) Its hard to keep track of all the issues so I will probably end up
>contradicting myself if I haven't already!

With the discussion becoming three-sided, expect me to find me
contradicting myself one of these days as well ;-)

>>There is not necessarily a concept of homology in vicariance biogeography.
>>The application of the term "homology" to matters biogeographical is purely
>>restricted to CGH Panbiogeography.
>ok. In which case there is no necessary biogeographic relationship to be
>derived from a biological analysis of relationships. Craw (1988) made the
>following observation: "Approaches to biogeography based on attempts to
>demonstrate distributional congruence are purely descriptive narrative
>approaches whether in the form presented above or bolstered by cladistic
>analysis of the taxa. They are not examples of analytical biogeography
>because geographic distribution is taken as read and not subjected to the
>equivalent of character analysis in form systematics to determine what
>aspects of geographic distributions of taxa are homologies.

I think we've been over this earlier in a thread on biogeographic
homology... My view is that the term homology is inapplicable outside the
context of phylogenetic analysis strictly.
Craw draws heavily on an analogy between phylogenetic and biogeographic
analysis. Granting the analogy, what would be the equivalent of  character
analysis? In phylogenetic analysis, the purpose of a character analysis is
to draw up a (putative) character state tree, which is then coded and
inserted in a datamatrix. Would not the equivalent of that be to find a
taxon-tree and use that as basis for the biogeographic analysis?

>>Accepting this usage, then, in vicariance biogeography we can observe
>>commonality when distributions (partly) overlap, in CGH Panbiogeography
>>when they overlap _and_ share the same baseline. Is that it?
>No. Distributions share the same baseline when their tracks cross the same
>baseline feature.

So - commonality is possible also when there is no overlap at all between
the distributions. It still seems to me that a major conclusion (my view!)
formulated on p. 161 "the major biogeographical regions are (...) the
modern ocean basins" is already implicit in this method...
The main question in my mind is why CGH Panbiogeography offers such a
limited choice of baseline features? Why not let the distribution patterns
speak more for themselves? If several patterns agree on a
Madagascar-Australia relation (e.g., fig. 1-9 A, 3-3, 3-5), - why on earth
"homologize" these with patterns involving India or Borneo (fig. 3-11)?

>>>>he criticism is invalid. It may be John Grehan's view that
>>>>explaining the distributions of individual taxa is not a useful enterprise
>>>No it is not.
>>Not your opinion, or not a useful enterprise?
>Not my opinion

That means that according to John Grehan  such explanations are  not "all
that informative at all - let alone 'coherent.'", but nevertheless
explaining the distributions of individual taxa is "a useful enterprise"?
I'm baffled.

>>I may have difficulties in defining "forcing", but I apply the term because
>>(as has been established in earlier posts), CGH Panbiogeography accepts
>>only 5 different types of trans-oceanic baselines (corresponding to the
>>major ocean basins), thus forcing all possible transoceanic patterns in
>>only 5 pigeonholes.
>When it comes to assigning ocean baselines there are indeed only five ocean
>basins. Baselines are not confined to ocean basins however.

It must be again an effect of the CGH Panbiogeography book concentrating on
only some parts of Panbiogeography as currently being performed - all the
examples in the book show ocean baselines, with the exception, perhaps, of
fig. 6-12, which shows global biogeographic track relationships based on
Croizat (1958). But then, I may be confused about the difference between a
generalized track and a baseline.

>>Seriously - what strikes me when reading Croizat is the stubborn refusal to
>>accept current maps as relevant for the explanation of distribution
>>patterns. A world of difference with the straitjacket of (current) ocean
>>basins imposed on biogeographical data in CGH Panbiogeography.
>Croizat used ocean basins to define global homology patterns so this does
>not lie just with CGH panbiogeography.

But Croizat had no problems in drawing different lines across the same
ocean (again, see fig. 6-12). In contrast, the comparable results of the
CGH analysis (fig. 6-13) show far less detail - with completely
undifferentiated ocean basins.

>>And the references?
Thanks - even if I don't agree it will make an interesting read.


P. Hovenkamp
Nationaal Herbarium Nederland - Leiden
PO Box 9514
2300 RA  Leiden
The Netherlands
hovenkamp at

More information about the Taxacom mailing list