Place in taxonomy

[Robert Mesibov]

I'm grateful for Tom's elaboration, because it spells out so clearly why
he's reluctant to use spatial data in constructing a phylogenetic
hypothesis. Let me try restating his view:

1. Geographical proximity doesn't correlate neatly with closeness of
relationship by descent. Two closely related taxa can live well apart
(dispersal) and two unrelated taxa can live close together (think of
Wallace's Line).

2. Spatial data aren't "massive" enough to assist the kind of statistical
analysis you need to pull a phylogenetic signal out of a blizzard of noise.
For example, 3 taxa XYZ living in 2 areas A & B gives you only 2 spatial
character states (A,B) to help in building the XYZ tree.

These are valid arguments in a general sense, but to apply them generally
is a mistake. There are numerous cases where geographical proximity DOES
correlate well with closeness of relationship, and in an analysis of
numerous taxa living in a biogeographically well-divided region, knowing
where things live can help you to discount (read: "judge to be less
plausible") trees which do justice to the character evidence but are
biogeographically unlikely.

Spatial information can be examined, organised and applied (as character
evidence is examined, organised and applied) while trees are under
construction. To use that information only after a tree has been
hypothesised is not to make the best use of all available evidence.

I may be reading too much into Tom's reply, but "given current
understandings about geography" is a worry. The biogeography of areas isn't
the same as the biogeography of taxa. In cladistic biogeography, a taxon
area cladogram is a shaky guess about the spatial history of a taxon. Taxon
area cladograms can be combined to a give an even shakier guess about the
spatial history of areas, but to "back-apply" that result to histories of
individual taxa is very dubious science.

It seems to me the issue is more methodological than philosophical. Tom
knows that there's phylogenetic evidence in spatial information, but he
doesn't yet have a good method for extracting that evidence. He shouldn't
claim that the evidence doesn't exist!
---
Dr Robert Mesibov
Research Associate, Queen Victoria Museum and Art Gallery
Home contact: PO Box 101, Penguin, Tasmania, Australia 7316
(03) 6437 1195; 61 3 6437 1195