More on reticulation

[Les Kaufman]

After dealing with haplochromine cichlids for some years, I'm forced to
conclude that most plants and some animals are not that different.
We've got a few discrete pieces of evidence for big-time reticulation
(hybridization between the Lake Edward and Lake Victoria sub-lineages)
but lots of fog everywhere else.  There is even a genus, Astatotilapia,
that encompasses a good clade of plesiomorph taxa that live in rivers
(being revised now by a colleague) and a huge block of what look like
primitive survivors of the early days of the radiation, all lumped
together arbitrarily because there's as yet no reliable way to tell them
apart, either with their genes up or down.

Astronomers have been able to cope with missing mass for many decades,
so we ought to be able to draw a chalk circle around the foggy spots and
say "there be mystery" to entice our students and occupy ourselves from
time to time.  The important thing to determine is not that the fog is
impenetrable right now- we know that- but that it is real and not imagined.

Curtis Clark wrote:
>
> At 03:02 PM 2/7/02, STEPHEN MANNING wrote:
> >>That's your call, of course, but most botanists I know won't reject a
> >>technique just because it's not infallible.
> >
> >I don't either, but could it be improved?  It's not my field, but I
> >sometimes wonder whether a statistician could come up with, or has come up
> >with, a mathematical model which would take into account reticulation in
> >such a realistic way that it would often alter, based on probable
> >paleopolyploidy, what is determined to be the most parsimonious
> >cladogram.  If about 50% of plant species are polyploid, this seems a
> >major concern especially when analyzing relationships of species within
> >genera, or genera within families (as both are subject to hybridization
> >events).
>
> Back when Mike Donoghue was at San Diego State, one of his grad students
> presented a paper making the case that reticulation could not be
> distinguished from homoplasy, i.e., that any character incongruence could
> be explained as parsimoniously by reticulation as by homoplasy. Honestly, I
> didn't follow the math, but it seems intuitively clear, especially if you
> accept things like horizontal gene transfer. And Verne Grant stated
> somewhere near the end of the second edition of Plant Speciation that we
> could never understand the phylogeny of flowering plants because of the
> high degree of reticulation.
>
> Nevertheless, there is a surprising amount of agreement between molecular
> and morphological trees for large parts of the flowering plants. Some of
> the problematic areas might represent reticulation, or not. But either
> Grant was for the most part wrong, or the evidence has seriously misled us.
>
> Polyploidy is a "red herring". Despite the claims of Stebbins to the
> contrary, autoploidy is not uncommon, so that a paleopolyploid doesn't
> always represent hybridization. Second, we now know of several methods of
> homoploid hybrid speciation. And modern alloploids often involve fairly
> close relatives (in the grand scheme of things); if alloploidy in the past
> behaved the same way, we would not necessarily be able to detect it ten
> million years later even if it wore a name tag saying "Hi, I'm reticulate."
>
> --
> Curtis Clark                  http://www.csupomona.edu/~jcclark/
> Biological Sciences Department             Voice: (909) 869-4062
> California State Polytechnic University      FAX: (909) 869-4078
> Pomona CA 91768-4032  USA                  jcclark at csupomona.edu

--
Les Kaufman
Biology Department
Boston University
5 Cummington St.
Boston, MA 02215
lesk at bu.edu
617-353-5560 office
617-353-6965 lab
617-353-6340 fax

and

BUMP
7 MBL St.
Woods Hole, MA 02543
508-289-7579 office
508-289-7950 fax