Phylogenetic evidence

[Tom DiBenedetto]

 Robert Mesibov wrote:

> Haeckel's term 'phylogeny' is now more than 130 years old and it seems to
> gain new shades of meaning each time someone sniffs out history in a data
> set for the first time. It has a core meaning, though. A phylogeny is a
> history of evolutionary events.

I find this definition to be overly broad, and thus taking us in the wrong direction. This
definition seems to indicate that phylogenetics is equal to the combined disciplines of
systematics and evolutionary biology. Phylogeny has always referred to the historical
relationships between species and groups of species (from the Greek - "tribe"-
"origin"). Phylogenetics is thus equal to the discipline of historical systematics - the
enterprise of erecting classifications of species based strictly on their historical
relationships.

> OK, so what kind of evidence could be used to erect a phylogenetic
> hypothesis, i.e. a guess at a phylogeny? Anything that contains traces of
> the history of the evolutionary events in question: fossils, biogeography,
> morphological character-states, molecular character-states, host-parasite
> relationships, etc etc.

And the upshot is that we do not necessarily look with equal interest at any data that
touches on the history of evolutionary events, but rather we focus on data whose
manifestations are directly reflective of the history of lineage diversification - e.g. the
distribution amongst species of particular heritable character states.

> Biogeographic evidence may be mutable and "blurry" at higher taxonomic
> levels but it does indeed contain traces of history. Every gene, every
> organism, every evolutionary event has (or had) an address. It often isn't
> easy to use biogeographical evidence but it is REAL evidence nevertheless.

Once again - it is REAL information, but I am not convinced that it can be considered
_evidence_ of relationship. Obviously every species has had an address. But for an
address to be evidence of relationship one needs to be able to say that TWO
species, with the same address, or with adjacent addresses or with addresses with
some definable relationship to each other are, on the basis of that information,
related to each other in some definable sense. For example - does having the same
address give you reason to propose that two species are sister-groups? Or does
having adjacent addresses indicate that? It is not enough to say "well, sometimes it
might" - because that doesn't advance our understanding one bit. That would not
make addresses evidence of anything.

> I gave 3 off-the-top-of-my-stack-of-reprints examples of using
> biogeographical evidence in an earlier post. The first one concerned
> reticulation evolution in a Tasmanian conifer genus. The evidence comes
> from total-genome DNA fragment analysis and from geographical mapping: the
> putative hybrid occurs only where the putative parents are sympatric.
> Imagine for a moment that the putative hybrid had instead been found in New
> Zealand. New hypothesis, maybe?

As to the mechanisms of how the evolutionary event happened, yes - maybe. But a
new hypothesis as to the relationships - no, not necessarily. If the character evidence
indicated that a newly discovered New Zealand species were actually a hybrid of two
Tasman species, would you advocate denying that conclusion simply because you
could not immediately envision how the hybrid got to New Zealand?

> Turns out that eucalypt cpDNA haplotype families aren't faithful to
> morphospecies lineages. Instead, they're endemic to particular AREAS within
> Tasmania, regardless of what eucalypt morphospecies happen to be living
> there and housing them. The authors see this as phylogenetic evidence of
> massive hybridisation in the past, perhaps when the eucalypt ancestors were
> crowded together in Pleistocene glacial refuges. Has Tom got an
> interpretation (a phylogenetic hypothesis) for these eucalypts that ignores
> the geographical evidence?

Without reading the papers, off the top of my head I would wonder why the
geography is seen as crucial to understanding the character evidence. I am not
questioning the conclusion of massive hybridization - or the fact that this might have
happened in a particular area. But had it happened differently - had the hybridization
not necessarily been geographically restricted - we might well have found the same
character evidence and reached the same conclusion. What does the geography add
to the understanding of relationships? N.B. I recognize that geography might well add
to the details of the evolutionary narrative - but my concern is with the question of
what is to be considered valid evidence of relationship - phylogeny.

Tom diBenedetto