More on paraphyletic species

Tom DiBenedetto tdib at UMICH.EDU
Tue Jun 17 09:55:33 CDT 1997

On Mon, 16 Jun 1997 17:28:08 -0600, Harvey E. Ballard, Jr. wrote:

>In my study of the Viola canadensis complex, and in the published
>literature that I was referring to that included multiple populations of
>widespread species and geographically restricted species, populations of
>the widespread species were phenotypically indistinguishable.  Assignment
>of these populations to the same species made no reference to a biological
>species concept.  In the Viola canadensis complex, the other species (the
>narrow endemics) differ from V. canadensis by a number of floral and
>vegetative features as well as a distinct preference for rock as opposed to
>soil substrate.  So the problem here is not one of a biological species
>concept in conflict with a historical context but one of phenotypically
>"the same" set of populations that are genetically divergent
>and--considering the embedded position of the endemics--paraphyletic.

The character evidence in this case seems, from how you describe the
situation, to lead to an unresolved polytomy; is this correct? If so,
a cladistic analysis would specify just that, concluding that the
relationships between the taxa in the complex could not be further
specified, given the evidence accessable through character analysis.
The point is that *based on character evidence*, there is no basis
for concluding that one "species" is paraphyletic; I dont see how you
are justified to make conclusions about the ancestor from an
unresolved node. You consider the widespread species to be
paraphyletic because you claim to know that it has budded off these
narrow endemics, and you base this claim on,,,,,well, I dont know on
what basis you claim it, perhaps because it is a scenario that comes
readily to mind? Is there not the possibility that the ancestral
population to the complex was a narrow endemic, that perhaps it was
widely dispersed, colonizing only rocky habitats, and that
subsequently one of the endemic populations diverged to give rise to
a form which became the widespread species?
        But lets imagine that your scenario is correct for a moment;
that there have been multiple independant colonizations of rocky
habitats by local populations of the widespread species. Given our
wonderful time-machine, we could perhaps observe the history of the
complex; we could see the historical relationships of one local
population to another, we could witness the colonization of a
particular rocky habitat by one of the local populations and the rise
of an endemic. I dont think that it is a uniquely cladistic
assumption that there really is a precise history that one can hope
to reconstruct. The problem of course, is that without the
time-machine we are constrained to reconstruct only those parts of
the history which have left evidence in the pattern of character
distributions. By regarding the widespread species as a single entity
(justified for now, given the state of our knowledge) we may be stuck
with a vision of a paraphyletic assemblage budding off isolated
endemics, but this is not necesarily an accurate statement of the
true history, rather an artifact of our lack of knowledge. The
accumulation of an ever more precise undersanding of lower level
relationships, leading to the deconstruction of paraphyletic
assemblages, is the basic work of historical analysis. Paraphyly is
inherintly a symptom of a lack of adequate information. I am sure
that there are many cases in which this will represent an absolute
barrier; historical information can be overwritten and gone forever.
We will never be able to discover the exact history of everything.
But that doesnt mean that there isnt a precise history which actually
happened, or that paraphyly is somehow a natural fact.

> I can't justify, any better than anyone else might, subdividing
>phenotypically indistinguishable sets of populations to make them
>monophyletic units to be called species, to "relieve" Viola canadensis of

I wouldnt ask you to subdivide them without evidence, but I would
urge you to seek out the evidence. That is the work of systematics.
Paraphyly is a call for relief!

>  But I'm talking about a more involved issue:
>when populations appear to belong to precisely the same taxon
>phenotypically, yet when analyzed cladistically using other types of data
>(molecular data) one arrives at a paraphyletic arrangement of some
>populations of the SAME taxon within which are embedded populations of
>other phenotypically divergent taxa (each forming a monophyletic group of
>populations that are phenotypically indistinguishable among themselves).

Ah good,,,so there might be historical information after all. I guess
it is time to redefine your taxa in light of the molecular evidence.
If it seriously contradicts the phenotypic evidence you will have to
find a way to resolve the dispute (default strategy is always of
course,,ever more evidence).  But molecular evidence is not
inherintly weaker or stronger than phenotypic evidence.

>Two sources of data "collide" but provide potential illumination when
>considered in an evolutionary context.  I confess that I don't see where I
>am neglecting the "historical" context.  On the contrary, the historical
>(or at least the geological and phylogenetic) context suggests derivation
>of the monophyletic sets of populations from different marginal populations
>of the more widespread and phenotypically diagnosable species.

Fine. And if you can distinguish the marginal populations from the
other populations of the widespread species, then you will close in
on a precise history of the complex.

> Which set
>of data is right regarding monophyly--or don't I have to choose?  I suggest
>that it's a larger issue than just choosing monophyly over paraphyly, or
>dumping a biological species concept or another that is not strictly
>cladistic in interpretation.

I 'm not sure I understand the question. Is the evidence actually
It sounds like you are discovering cladistic structure within the
widespread species through other-than-phenotypic evidence. And that
maybe the endemics can be associated with the varying parts of this
structure. Sounds like you are on the right track,,,except that you
resist the notion of assigning equal rank to an endemic (species) and
its sister-population (one of the marginal pops of the widespread).
You insist on ranking  the widespread as a whole as a species. So
your ranking is based on similarity (actually only phenotypic
similarity) and not on the actual hierarchical structure. Perhaps you
see the taxic hierarchy, but apply the rank in a way which dosnt
respect that. It is this decision of yours which *creates* paraphyly.

[re. un-recognizing paraphyletic taxa]
> Why not take a conservative approach and leave them where they
>were originally placed until further data clarifies their position (or
>doesn't, as the case may be).

I agree. I dont support making formal changes in particular taxonomic
issues except by those who have studied all aspects of the question
(re-analyzed all data previously accumulated, added their own,
arrived at a "state of the art" assessment of all the evidence
bearing on the issue).

>In this instance,
>a group of phenotypically indistinguishable units is rendered paraphyletic
>by other groups of indistinguishable units.  We could make up names for
>each of the sets of indistinguishable units, or turn to microcharacters to
>try to resolve differences.


>  But assuming that we find no other phenotypic
>features that allow us to reassign the paraphyletic set to a number of
>different monophyletic assemblages such that every set is monophyletic, how
>do we deal with the situation?

I agree with the concern,,,lacking evidence we cant do much....that
doesnt make the paraphyletic assemblage 'real" though,,,its just a
lack of evidence. Look at sequences.

> All the evidence suggests a particular
>evolutionary event: speciation around the periphery of a widespread taxon
>represented by a series of populations across its range.  Cladistically
>this creates a discomforting pattern that isn't rectifiable on a molecular
>phylogenetic level with phenotype.  What now? My argument is that this
>performance receives a frequent encore throughout geologic history.  How
>will a usable PHENOTYPIC nomenclature be developed to accomodate this as
>well as "clean" and balanced phylogenetic splits that result in two sets of
>populations that are essentially indistinguishable amongst themselves and
>divergent phenotypically from other related populations.

This is why I (sort of) accused you of not being grounded in the
historical framework. My bottom line is to have a classification
which reflects our conclusions regarding history, and if that is
messy from a strictly phenotypic perspective,,well too bad,,thats
life :) .

More information about the Taxacom mailing list