Biogeoclimatic zones and homology

[Don McAllister]

John Grehan wrote:

> The biogeoclimatic zones described by Ron Kaneen are indeed arbitrary
> biogeographic units. The units are defined by a selected feature enclosed by
> a "line" representing the geographic limits of that feature (or some kind of
> approximation of that limit). Such area concepts are very popular. Whether
> they are worth the effort I don't know.
>
> An biogeoclimatic area may be defined by a particular feature
> such as the range of a particular plant group or combination of plants, and
> for that feature the range is certainly recognizable, but the area itself has,
> in my opinion, any "natural' homology in terms of the orgin and evolution of
> distributions.

I agree with Grehan to a considerable degree.  But is generally ignored is that
the starting geographic unit for which one assembles the biological data is
usually a country, continent, island, etc., that has no sound scientific
rationale.  It is merely a convenient pigeon-hole.  The problem is that it
confines ones hypotheses to that pigeon-hole unit.

The advantage of a global equal-area grid is in part that that the  geographic
units are equal in area, hence need no area-correction in making comparisons. But
a more fundamental one is that the grid can be used in a wide-open exploration of
where the biological-geographic areas of interest are really located.  Proceeding
in that fashion, whether one is examining endemism, cladistic linkages, geological
history, one is not tied, unconsciously or not, to some arbitrary geographic
designation.

Thus biogeographic areas of interest can be recognized defined thorugh use of an
open-minded approach.

Equal-area grid data can also be used for open multi-variate principal components
(PCAs) and other multivariate analyses (though I have not done so), as well as
relatively simple mathematical treatments.  The clusterings produced by PCAs are
not determined by any a priori hypotheses, but are generated by the nature of the
data.

With the equal-area grid approach another limited factor becomes apparent - the
nature and quality of the data.  Much of our geographic data is gathered randomly
and opportunistically.  Terrestrial distribution data often mimics the road
network! Our coral fish data showed often that some Pacific island reefs simply
had not been properly sampled, producing what are probably artificial gaps in
known distribution patterns.  This is not a flaw of grid analysis.  It is the fact
that biological surveys are underfunded and poorly thought out.  Good
biogeographic analyses require good biological data, gathered in a regular and
planned manner, not by hazzard.

The other basic limitation is the generation of sound taxonomies. This can be
thought of at two levels.  Firstly what are the valid species, the units which
usually have the most objective biological reality?  Research will reveal what
species there really are, what names should be regarded as synonyms, what are
merely subspecies or local varieties.

The second level is that of relationships, whether it is the establishment of
genera, families, orders, etc., or the identifcation of clades - both hopefully
through the examination of several suites of characters so that the relationship
of individual units do not rest feebly on one characteristic only to have the
classification/cladogram swing wildly when it is based on another character set.
The consideration of information in higher taxonomic units or clades is valuable
and deserves more attention.

don
Don McAllister