A biological continuum
george_garrity at MERCK.COM
Fri Mar 31 11:18:23 CST 1995
Reply to: RE>>A biological continuum
I for one would like to see the discussion continue. A number of relevant
points have been raised that warrant serious discussion and this forum is one
of the few in which perspectives from different disciplines are available
(e.g. botanists, zoologists, paleontolgists and at least one microbiologist).
As our studies progress, I'm witnessing something that doesn't fit into the
classical framework of Linnean taxonomy. But then again, should I be
surprised ? Do the models that we use adequately explain what is happening ?
Are there better models ? Are heirarchical models appropriate ? What biases
exist in the different models and how does one uncover them ? What happens
when we move from retrospective to prospective taxonomy (e.g. surveys of
biodiversity) ? More importantly how do we begin to consolodate all of this
information into a more coherent view and use this infomation intellegently
and productively ?
Natural Products Discovery
Merck Research Labs
Date: 3/31/95 9:33 AM
To: George Garrity
From: DARBYSHIRES at NCCCOT.AGR.CA
>> A) Curtis, you're still stuck on vertebrates.
>A strange "accusation". While I once was a zoologist (for as long as
>it took to get a B.S.), I am now known among my colleagues for
>criticizing "vertebrate-centric" evolutionary thinking. I base
>my statements on evidence from plants. (Or were you implying that my
>plants have backbones?)
My point is that non-sexual reproduction is only the "rare
exception" you make it out to be among vertebrates. You began
this thread by pointing out that continua are not found in
*eukaryotes* (as opposed to prokaryotes), which as a group are
considerably larger than vertebrates and vascular plants put
together. Perhaps you, or someone else, might care to make some
comments about the eukaryote kingdoms of protists and fungi at
this point. My point being that you made a sweeping
generalization based on your experience with a *small* subgroup
of organisms with a rather depauperate set of mechanisms for
> > B) I do not know what you mean by "cladistics...works well".
> > Cladistics as practiced today often produces one (or many
> > more) *hypothetical* phylogenies that "in many cases"
> > appeals to the subjective conceptions of the worker. Even
> > if one accepts the theory of cladistics (presumably we are
> > talking about the Hennig derivative procedures here, which
> > do not permit the consideration of the continuous nature of
> > organismal phylogeny through time)
>This is exactly my point. It seems by your post that you do not like
>cladistics; opinion noted. Nevertheless, a system that by your admission
>does not (see quote above) is nevertheless capable of producing congruent
>phylogenies of many groups from different sorts of information. The
>cases where cladistic methodology doesn't work in my view provide an
>estimate of the extent to which evolution has not taken a branching
>topology. Cladistics (to me at least) is a tool, neither God nor Satan,
>and it shows me countless examples of groups without reticulation, and
>groups in which reticulation is a minor feature.
Don't jump to conclusions: I certainly do "like cladistics" (your
words), in theory; it is an interesting tool for developing
hypotheses (not exposing truths) but limited by the worker's
subjectivity and assumptions. There are many aspects to
evolution that clearly do not fit with the methodologies of
popular cladistics, e.g.: evolution is not necessarily always
cladistic or parsimonious. As a procedure, cladistics cannot
detect intermediates, or any assumption violations. Characters
that are variable within a taxon are eliminated from an analysis
because they are uninformative. There are just as many cases
where cladistic analysis of different data sets for the same
organisms result in different topologies as they result in
similar topologies. What are these "countless examples" that
have been studied using different equivalent data sets (another
rather extreme generalizing statement).
> > one sees what one wants in the results. The presence of
> > reticulation cannot be determined by popular cladistic
> > procedures, in fact one starts with the assumption that
> > reticulation does not exist.
>Reticulation causes cladistic techniques to fail. When they work, this
>is evidence for the reduced importance of reticulation.
I still don't know what your criteria are for judging that
"cladistics works". I maintain that it will not show
reticulation because the starting premise is that there is no
reticulation. Just because the results of cladistic analysis of
different data sets are not congruent does not mean that
reticulation is the culprit.
> > C) Reticulation is very real, even in vertebrates; to
> > mention two cases: the complex of salamanders in the
> > genus _Ambystoma_ from eastern North America. This group
> > has been studied intensively since the early 1960s only to
> > reveal more and more complexity to the reticulation. Four
> > diploid species are known to be hybridizing and exchanging
> > genes (A. laterale, A. jeffersonianum, A. texanum, A.
> > tigrinum) and other species are suspected to be involved.
> > Diploid, triploid, tetraploid and even pentaploids are
> > known; polyploids being mostly female.
>More zoology stuff, but I've left this in because it supports my point.
>This is a notable example *precisely because it is notable*. If
>reticulation were the rule and branching phylogeny the exception,
>this would not merit mention, and we would focus on the unusual
>groups that *don't* show reticulation.
This was one, admittedly extreme, example, but the point is that
you originally said "eukaryotes", you used the term "sex" in a
`traditional' sense and unless each case is studied carefully in
depth, it is unwise to assume either a continuum or no continuum.
Continua exist in all groups through time and space; reticulation
(in many forms and many taxa at different) is real.
> > Also among the most
> > heavily studied organisms are cereal grasses and their
> > relatives. Wheat, barley and rye belong to the Triticeae
> > tribe, in which a vast array of `intergeneric' hybrids have
> > been documented where two or more distinct genomes have
> > combined in hybrids to give an incredible taxonomic tangle
> > of intergrading species and genera.
>Sure. I lecture about this all the time. But any subsistence farmer
>from central Asia can tell the difference between einkorn and emmer.
>Just because there is reticulation here, that is not evidence of continuous
>variation. And in the example of cultivated wheat, the chromosome
>numbers of the polyploid series are discrete, with no true-breeding
Any American Kennel Club member can tell the difference between a
terrier and a chow chow. Maybe even botanists can be trained to
do this. :-)
Have you looked at the _Elymus trachycaulus_ complex throughout
the northern hemisphere? A lot of botanists would appreciate it
if you would point out the discontinuities in this species
complex after studying it (using cladistic methods or others)
throughout it's range. As I said, there are hiatuses in some
places, but in others there are continua.
> >Are these plants asexual F1 populations; no segregation at F2
> >generation, no chance for backcrossing? Are these plants which
> >normally only reproduce through chasmogamous, outcrossing
>No, no, no, yes. "Homoploid hybrid speciation" = Origin of new diploid
>true-breeding species from hybrids between other species. The group
>I work with is Encelia of the Asteraceae; these are obligate outcrossers,
>F1s are fully fertile. Backcrossing in cultivation is simple and backcrosses
>are fertile and vigorous, but backcrossing in nature is rare for ecological
>reasons that we are still trying to figure out. Hybrid species appear to
>have resulted from crossing among F1s, F2s, and so on, with selective
>elimination of plants tending toward either parent's morphology. You
Sounds like a specialized case to me. Probably _Encelia_ hybrids
are an exception rather than a rule. :-)
>might be surprised to know that one of the lines of evidence supporting
>the hybrid origin is cladistic: co-occurrence of autapomorphies of the
>From the little information you provide in this example this
could be interpreted in other ways. Other data have helped
interpret and resolve conflicting cladistic hypotheses generated
by popular methodologies (that's fine). Cladistic methodology
`proved' nothing, but simply indicated a number of conflicting
and unresolvable hypotheses.
> >Introgression does not result in new species, agreed, it results
> >in continua.
>No it does not. The mere fact that we can recognize a situation as
>introgression is evidence that the parent species are distinct in at
>least part of their range. Yes, there are intermediates, but the bulk
>of variation follows the parental means.
Perhaps we are taking different views on the geometry of
continua. My concept of continua includes a spacial component,
yours does not seem to.
> >OK, in some cases intermediates are not important "statistically
> >or biologically", but do not restricting sampling to specific
> >cases or in time or in space or in numbers, so that few
> >intermediates are detected. To say it again, the extent of
> >variation between forms is continuously variable, from uniform
> >continua to wide hiatuses. To deny the existence/significance of
> >continua is to deny an important aspect of biological variation.
>Let me try my hand at ASCII graphics:
What are your end points (or modes) here. Have you included all
eukaryotes in this graph? Is this the graph that one gets from
looking at "Homoploid hybrid speciation" events. If so it seems
to be one lump short of a full load :-) Sorry, another cheap
shot, but I couldn't resist. Seriously now, do you see this
pattern between *all* adjacent modes?
>You stress the presence of intermediates. I stress the bimodality.
>I do not deny the presence of intermediates. You seemingly deny the
>importance of the bimodality.
There may lie our problem. We both see each other as being
altogether too dogmatic about the points we are trying to get
across. The concept of bimodality (or multimodality) will
operate at many levels: population, subspecies, species, genus
(or similar artificial ranking) -- do I detect another form of
continuum? Does your concept of bimodality operate through
multiple dimensions or is it simply a frequency at a particular
point in time and space. Obviously it would be a bit
hypocritical to work as a taxonomist without recognizing the
existence of some sort of discontinuities, partial or complete,
somewhere. Without meaning to criticize any reader of TAXACOM,
especially you Curtis, I note that it is easy to detect
"bimodality" using a limited number of samples. This is not so
much of a problem now as it was for botanists of the last
> >Maybe the difficulty is that your work is done on California
> >plants (?).
I think not, although it might seem like it. The point is that
my concept of continua, as alluded to above, is one that tends to
consider large geographic areas (as well as evolutionary time --
did those 12 synapomorphies defining the clade arise in one
> > I suggest that you might try to examine a
> >circumboreal species complex (animal or plant) *throughout it's
>I could suggest that *you* examine a case that supports *my* views,
>but instead I'll agree to disagree.
Well a specific example has now been alluded to (_Encelia_ in
California, right?), I will try to look it up, but how can I be
sure that it's not a special case exception? :-) Thank God we
"disagree", otherwise the discussion would flounder to a rather
boring ending. Thanks for the discussion Curtis!
Too bad everybody else on TAXACOM is ignoring this thread and
jabbering endlessly about ZPG, maybe it means that we are both
being trivial and uninteresting (if not long-winded).
Because I like quotations so much (especially when they apply to
me) and because others seem to have enjoy the Veldkamp quote, I
leave you with another of my favourites:
... I have long thought that too much systematic work
[and] description somehow blunts the faculties, ...
More information about the Taxacom