Explanations in biogeography

John Grehan jrg13 at PSU.EDU
Mon Oct 4 02:15:24 CDT 1999

Stuart Poss wrote:

>I don't know any biologist who would argue that an an organism's geographic
>location is not informative.

But in historical biogeography the only research framework that I am aware
of that makes geographic location (and spatial relationship) and explicit
component of the analysis is panbiogeography. Dispersalists focus on
dispersal (migration), and vicariance cladists focus on biology (as

 Indeed, to have knowledge of any organism one must
>be able to observe it somewhere.  Likewise, the same is true of time, and hence
>fossil evidence is particularly important in biogeography.

Fossils give the minimum age of fossilization, so are important for what they
contribute in this respect, and in providing additional locations.

Time and space are
>not solely the provenance of panbiogeographers.  Others have been known to
>and observe the effects of both.

Perhaps I am prone to some exageration. In historical bigoeography
panbiogeography is
the only research program that defines geographic information as its mode
of analysis.
Yes, individuals may engage in various forms of spatial correlation, in
which case
they are often using techniques that also exemplify panbiogeography. These
may be panbiogeographers without knowing it.

>In a previous post, you note that if a track is correlated with a vicariance
>event, then you can make historical predictions.

I apologize for not making myself clear. I was not correlating tracks with
vicariance events (which are really just hypotheses about the past), but with
geological/geomorphological structures.

 I would not disagree, so long
>as you have information that this is in fact true.

Geological/geomophic structures are empirical facts, but their historical
may be open to broader alternative interpretations.

However, my point is that
>simply making the assertion that a track drawing an association between
>two areas
>exists does not provide an explanation for a disjunct distribution.


 This is true
>because a wide number of possible interpretations of how such an
>association, if
>real, could have arisen.  One must be able to choose among them to arrive at an
>explanation.  A given disjuction could have resulted from vicariance, but it is
>also clear that disjunct distributions in many cases have not arisen from

Agreed. Panbiogeographic analysis is focused on the spatial relationships
between different distributions. Chosing between vicariance and dispersal
is not as central to panbiogeography (in my opinion at least) as it is in
dispersalist or vicariance cladistics.

>You speak of "spatial structure", but in many cases the relative duration
>of such
>"structure" must be demonstrated before it can be assumed.

Spatial structure is as it is at the time of analysis. Agreed, its
structure may have changed
over time, but the track is about spatial geometry, not about identical
The fact that there are disjunctions between localities linked by the track
implies that there have been changes in the distribution over space and

In my view, simply
>drawing tracks between disjoint distributions and making assertions about
>vicariance in the absence of information that would preclude extinction and/or
>multiple episodes of dispersal along the same (geometric/spatial) track
>does not
>provide a compeling case for discriminating between dispersal and vicariance as
>explanations for the disjunction.

This is a matter of individual judgement. If Stuart Poss and others feel
that the panbiogeographic method is deficient, they have the right to this
assessment. Others may feel differently.

>Your arguments will be more likely to be useful when the vagility of a given
>species is low and it is so fixed or adapated to a particular habitat that the
>probability of its dispersing elsewhere is extremely low.

We make some observations on such views in the panbiogeography book. Mayr
claimed that poor dispersers such as earthworms and primary freshwater fish
have totally (and I emphasise totally) different distributions from winged
birds and butterflies. We show that this is not an empirical truth. Haydon,
Crowther and Pianka similarly claimed that the biogeography of bats and
spiders confounds attempts to recover general patterns and responsible
processes. Again, whe sho that this is not an empirical truth.

  Certainly, the
>presence of caecilians on various Indian Ocean islands do provide good examples
>for disjunct distributions that seem explanable only through resort to plate
>tectonics.  However, one can not easily use the same tracks to explain the
>presence of certain species of the lancelet genus Epigonichthys at many of
>same islands.  One can draw the tracks, but their explanation are in all
>probability quite different.

Perhaps, perhaps not.

Generalization is not warranted.  No "pattern"
>exists, just a confounding of multiple "patterns".

There are a number of standard tracks that have been identified for the
world. These are recognizable in different groups.
>Since, I believe in earlier posts you noted an unwillingness to accept natural
>selection as the primary mechanism of evolution, it is not surprising that you
>find it necessary to adopt explanations of geographic distributions that do not
>involve natural selection or effects that do not befall upon individual

I agree, perhaps it is not suprising. On the other hand, it would be a
mistake to assume that
all practioners of panbiogeography feel about natural selection the way I do.

 I would argue that such an approach will always be futile even in
>cases where vicariance does explain the disjunction, because the information
>necessary to make the distinction among different possible "tracks" associating
>disjunct parts of a geographic distribution will come from the biology of the
>organisms and not the space they occupy in a purely geometric sense.   To see
>that this is true one need only observe that the notion of track geometry is
>insufficient to explain distributional limits.

But track analysis does allow one to determine the level of generality
about distribution limits and show that not only may different taxa with
different dispersal abilities share a common biogeographic limit, but such
limits may be spatially correlated with tectonic or other geological
structures. This is information that would not be obtained other than by
some spatial analysis.

If only spatial geometry were
>required to infer the origins of distributional patterns, how could one account
>for the inability of many organisms to occupy the "spot next door" even though
>they have disjunct distributions that are widely separated?

I don't know what the problem is here with respect to panbigoeography.

 Perhaps there is
>another part of the panbiogeographic algorithm that still needs to be specified
>to account for such important biogeographic phenomena.

Perhaps the node concept will be relevant.

John Grehan

More information about the Taxacom mailing list