natural units in biogeography

[P.Hovenkamp]

At 10:19 PM 26-04-99 -0400, John Grehan wrote:
>>At 08:09 AM 27-04-99 -0400, John Grehan wrote:
>>>Peter Hovenkamp's assertion that  "Natural classification" has no meaning
>>>in biogeography represents a fundamental difference between our respective
>>>views.
>>>
>>>I agree that evolution of biogeographic histories is not the same thing
>>>as evolution of organisms, but I do not see that more than one context
>>>may be applied to evolution.
>
>
>>You can always use the word "evolution", or "evolve", like in "the
>>evolution of the stars", but that does not mean that that all evolving
>>histories are isomorphic in the sense of showing a pattern of nested
>>hierarchical relationships.
>
>This I can agree with, but this does not translate as justification for
>a biological application of synopomorphy being the only possible
>foundation for a natural classification, and that natural classifications
>are not possible in biogeography.
Now that we systematists have some sort of agreement about what exactly is
"natural" in biological systematics (something having to do with
evolution), we might start again in biogeography. Of course, we then have
the restriction that "natural" can not be made to mean the same thing as in
systematics, but apart from that, the discussion is open.
Only, I seem to remember that this thread started with a question about the
application of synapomorphy in biogeographical classification. If that has
been dealt with, I'm happy.

>
>In terms of a natural classification for biogeography, the application of
>natural units in panbiogeography informs us about a shared history
>for the origin of different distributions with respect to a specific
>geological/geomorphic feature. If a distribution has, for example, a
>Pacific baseline, this homology statement allows the hypothesis that the
>origin of the distribution has something to do with the formation of the
>Pacific (it is a hypothesis just as a character individually identified as
>a synapomorphy is a hypotheses). The baseline allows this distribution to
>be compared with other distributions sharing the same baseline, and
>distinguished
>from thos that do not.
>
>Here the application of natural involves classification according to
>hypotheses of a historical relationship between a spatial biological pattern,
>and a spatial geological/geomorphic pattern. I would not purport to
>know anything about natural classifications of stars.
>
>
>I think the starting point for vicariance biogeography
>>is that if we see multiple sister-group relationships across a particular
>>line, we may be justified in thinking that that line is the location of a
>>vicariance event.
>
>Interestingly how does one know that a sister-group relationship
>crosses a particular "line"? If a taxon is present, for example, in
>the Netherlands, Japan, and Columbia (and assume that the biological
>relationships are Netherlands-Japan, and then Columbia) and that
>several other taxa share the same pattern, what is the spatial homology
>of the distribution? How does one determine what geography, out of
>all the possibilities, are involved if there is no natural classification
>criterion?
I'm a little confused about a few points in the discussion above. When you
say "distribution", you seem to be talking about patterns of taxa on any
level. I think there are two things to be distinguished there.
1. A species-taxon distribution is no evidence for a historical association
between its constituent parts. The alternative explanation, current or
recent contact allowing dispersal must be taken as equally likely.
Concrete: if a species is shared between Japan, The Netherlands, and
Colombia (and no other areas is what I suppose you mean) it seems obvious
to me that it was sent from Japan to The Netherlands by Von Siebold in the
early 19th century, and that the occurrence in Colombia is based on a
similar event, or on a misidentification.
2. A distribution of a taxon on a higher level may much more confidently be
considered evidence for vicariant speciation. Thus, "baselines" based on
genus distributions must be distinguished from those based on species
distributions. All genera except monotypic ones can be expected to have at
least one a bifurcation in their phylogeny, and thus there is potentially
at least one sister-group relationship hidden in each genus. If we cannot
correlate a number of such "vicariant" sister-group relationships to each
other, or to a recognized event in earth history, wee are getting somewhere.
If not, we may still be looking at freak events of dispersal. The
occurrence of a genus shared between The Netherlands, Japan and Colombia
only probably the result of two such events. I can't think of a genus that
shows this distribution, actually, which may be taken as an indication of
the improbability of two such events happening.


P. Hovenkamp
Rijksherbarium, Leiden
The Netherlands
hovenkamp at rhbcml.leidenuniv.nl
http://rulrhb.leidenuniv.nl/