Place in taxonomy (finish?)

[pierre deleporte]

In my view, Robert Mesibov brings some convincing examples of what I called
"Progress in approaches of limited scope", constituting "a useful first
step" for improving biogeography methods, at least for recent events.

Phylogeny inference requires the use of all relevant evidence, and
information for hybridization matters. Evidence for hybridization may
effectively come from spatial data, and from genetic data other than
classic sequence characters, however under some interpretative models
(e.g., hybrids should generally stand in sympatry with both parent species,
or hybrids should share that proportion of features with parent genomes).
And, if accepted as solid evidence, be used to correct "rough" phylogenetic
inferences.

Combined with a classic vicariance approach for instance, it would mean
that phylogenetic biogeographgic inferences should not be done without
first checking for (possibly spatial) evidence for hybridization.

Not a software package yet of course, but possibly a growing series of
logically consistent recommendations for interpretation of typical cases.

Thanks very much for the examples

Pierre
A 10:28 04/07/2002 +1000, Robert Mesibov wrote:
>Pierre Deleporte has asked how, specifically, you can extract historical
>evidence from distributional data. Unfortunately I have no Method (note the
>capital) and no software package (Geoclade?) to offer, but here are 3
>recent papers to think about:
>
>(1) The 3 conifers Athrotaxis cupressoides, A. laxifolia and A.
>selaginoides are all endemic to Tasmania. Each has its own set of
>morphological and ecological character-states, and each produces fertile
>pollen. It's a straightforward exercise to devise 3-species cladograms from
>character-states, and you can then fiddle with statistical tests to see
>which of the cladograms is most likely. All the cladograms are wrong.
>Genetic evidence (RAPD & SSCP analysis of total genomic DNA) shows clearly
>that laxifolia is a stable, interspecific hybrid of cupressoides and
>selaginoides, and selaginoides may well be the only pollen parent (Aust J
>Bot 48:753-758, 2000). Now let's expand on that "endemic to Tasmania".
>Mapping Athrotaxis distributions on a fine scale, you find that laxifolia
>occurs ONLY where cupressoides and selaginoides are sympatric. Here the
>best evidence for clarifying the phylogenetic relationships between 3
>species comes from genetic and spatial data, not from a taxon-by-character
>matrix.
>
>(2) In a fascinating study of grasshoppers in the genus Chitaura on
>Sulawesi (Biol J Linn Soc 72:373-390, 2001), 4 types of data were
>available: exact locations for samples from 60 sites, maximum parsimony
>cladograms for colour pattern, a mtDNA cladogram and a morphometric data
>set from males analysed using PCA, HANOVA and a Mantel test (with
>geographic distance). There are 10 named Chitaura species and the character
>data certainly support the conclusion that the Chitaura lineage has
>branched repeatedly over the past few million years. However, it is far
>from obvious how a conventional tree can be constructed which satisfies all
>the data. Instead the authors focus on spatial relationships (especially
>what happens at contact zones on the island) to illuminate tree "details".
>Read the paper to see how complicated the story is.
>
>(3) A preliminary analysis has recently been made of the connections
>between species range size/position and hypothesised phylogeny (Amer Nat
>155:419-434, 2000). In the groups studied, range size asymmetry was
>associated with relatively recent splits, suggesting that peripatric
>(allopatric) speciation was the split mechanism, and sympatry seems to have
>resulted from post-speciation dispersal. The hint here is that the geometry
>of a set of ranges contains phylogenetic information which would not be
>evident in a simple presence/absence summary of locations.
>
>In an earlier post I pointed out that spatial information becomes less
>important (more blurred) as you go up the taxonomic hierarchy. A reckless
>generalisation: spatial information is essential in studying populations,
>very useful for species, important for genera and helpful for families.
>This shouldn't be seen as a limitation which reduces the importance of
>"place" overall, any more than mtDNA analysis should be dismissed because
>it doesn't help us understand basal arthropod phylogeny.
>
>Dr Robert Mesibov
>Honorary Research Associate
>Queen Victoria Museum and Art Gallery
>Home contact: PO Box 101, Penguin, Tasmania, Australia 7316
>(03) 64371195; 61 3 64371195