Species Concept Question

[Richard Pyle]

Response to Jim, Part 2....

> >A deeper level of the question has to
> >do with the role of distribution patterns and historical nomenclature in
> >making nomenclatural decisions.  For example, if in Pattern 1
> population "C"
> >did not exist, would you be more inclined to treat them as
> separate species?
> >The gut-reaction answer is "Yes, because without evidence of gene flow
> >between two clearly distinct forms, one is to assume that they
> are distinct
> >species".  But if population "C" didn't exist, how would we ever know
> >whether the two forms were capable of freely interbreeding?
>
> Isn't this just a long-winded way of introducing the forbidden in polite
> society dinner time conversation topic ' What is a secies?'  (it is right
> up there with sex and politics...)

In a way, yes -- but as I tried to pin down later, I'm attacking it from a
rather specific angle -- one that involves consideration of the size and
shape of distributions outside the hybrid zone, as well as the bias of
sequence of population discovery.

> >  If, on the other hand, the pattern of distribution looked like this:
> >
> >   A       B       C       D       E
> >-----------------------------
> >XXX      XXX    XOX     OOO     OOOO
> >  XXX    XXXX     OXO     OOO    OOO
> >XXXX     XX     XOXO    OOOO    OOOO
> >
> >That is, no evidence of gene flow between the two populations in
> an area of
> >sympatry, then most of us would have no trouble arriving at the
> conclusion
> >that these are distinct species (i.e., given the opportunity to
> interbreed,
> >they do not).
>
> but if as you say the taxa are indisputably closely related this pattern
> does not rule out evidence of past geneflow, followed by some form of
> effective reproductive isolation...

Yeah, but....EVERY living thing on Earth shares some form of "past" gene
flow.  The logical (to me, anyway) litmus test for establishing a species
boundary is *future* gene flow.  Absent a population "C", we have no way of
predicting future gene flow with any semblance of confidence.  Given a
population "C" with apparently unrestricted geneflow between forms, we still
can't predict whether geneflow will eventually re-establish itself across
the broader populations, inclusive of A, B, D & E.  We have evidence of the
potential, at least -- so it seems to me a logical situation to implement
the subspecies boundary.  But in the example above, with well-established
sympatric populations of X & O, and no indication of multi-generational gene
flow, we have a farily compelling piee of evidence that non-geographic
barriers to reproduction are established.  In this case, the most
parsimonious prediction seems to me that gene flow between the two
populations has ceased, and will not likely resume ('though anything is
possible, of course) -- and therefore we should have confidence in
establishing a species boundary betweent the two.

> >But if, as I posited above, population "C" doesn't even exist (to our
> >knowledge), how do we assess whether the two morphotypes should
> be regarded
> >as distinct species (old question, I know -- and mostly rhetorical.
>
> that implies you already know the answer...  or assume that we do...  :)

No -- it actually implies that there is no answer -- at least no clear
consensus seems to have emerged from historical debates of this sort.

> >Also,
> >not really at the heart of what I'm after -- as I said, I can't
> quite figure
> >out how to articulate the real question I'm after).
>
> if it looks like a species, smells like a species, then call it a
> species...
> how difficult is that?   It is just a handle to an
> arbitrarily
> useful construct at the time...  :)

Again, if only the broader taxonomic world saw things the same way that you
and I do.  "A species what a taxonomist or community of taxonomists says it
is."  It's always worked for me (as a matter of fact, I first publicly
declared this to be my position when I gave a presentation about patterns of
hybridization in reef fishes, and someone asked me afterward why C.
flavissima & C. vrolikii should be regarded as separate species when they
clearly produce fertile hybrids when given the opportunity).  I only later
came to realize that it wasn't an original perspective...

But actually, this does relate to what I'm after.  I see the examples of
morphologically distinct populations across a large-scale patchy environment
with small areas of hybrid swarms to be a troublesome scenario for the
"species is what a taxonomist says it is" paradigm, because "reasonable
people can disagree". Or, maybe more relevantly, "a reasonable and normally
decisive taxonomist can remain undecisive".

Given the generic scenario I described:  Two broadly distributed and
unambiguosly distinct populations with very little intra-morphotype
variation except for relatively narrow zones of rampant hybridization; what
do people on this list feel is the most nokmenclaturally useful solution:

- Emphasize the differences with two species epithets;
- Emphasize the absence of non-geographic barriers to gene flow by treating
them as conspecifics;
- Emphasize both by invoking subspecific (or whatever trinomial) epithets?

I know that a lot depends on how different the two "parent" populations are
from each other, and proportionally how much of the total geography is
represented by the hybrid zones, and whether the populations seem to be
relatively stable, or are potentially volitile over evolutionary timescales,
etc., etc.  But part of the impetus for the post is that I and a colleague
looked at exactly the same pattern, and arrived at different conclusions (I
saw evidence supporting recognition at the species level, with
acknowledgement of a hybrid zone; he saw evidence for synonymizing the two
species).  One of the goals of this post is to develop a sense for which of
our perspectives resonates more with mainstream taxonomists.

> >Now, suppose throughout most of the relevant nomenclatural history,
> >population "C" was not known, and the two distinct morphotypes, each with
> >relatively broad distributions, were consistently treated as distinct
> >species by all researchers over many decades.  Then, someone discovers
> >population "C".  Are you now tempted to disrupt nomenclatural
> stability and
> >treat them as conspecifics, or would you prefer to maintain them as
> >distinct, and note a zone of hybridization.
>
> in pattern 1, why not, if that is what it look like?   In pattern 2 the
> decision may not be so obvious as what you appear to be describing is a
> gradual cline, and hybridism may or may not be involved or may
> only be part
> of the story...

Agreed on both counts.

> >  How would you approach the
> >situation differently if population "C" was discovered and known for many
> >years as a single, highly variable species, and only later was
> it realized
> >that the two endpoints of the variation spectrum were each represented as
> >broadly-distributed but allopatric populations?  Similarly, what if in
> >Pattern 2, only populations "A" and "E" were discovered
> initially, and named
> >as separate species, and then subsequent research revealed the
> existence of
> >a cline of populations?
>
> Surely the historical sequence of events should not make any difference
> once all the results are in?   Don't you just look at all the available
> data and make the decision regardless of what people had thought before?

Yes, in most cases.  But this is one important facet of the point of my
post.  Does a many-decade, highly stable taxonomic history nudge you in the
direction of maintaining separate species epithets, in cases where your
taxonomic barometer could go either way?  In other words, what role does the
preservation of nomenclatural stability play when you are faced with a
border-line case in your own personal algorithm for assessing the
species-boundary threshold.

> >As I re-read this note, I am still not satisfied that I have conveyed my
> >true question.
>
> Don't try *again*... Please!

It's Good, Jim.  GGOOOODDD! :-)

> >the ways that history and
> >observation of geographic distribution influence and bias our
> nomenclatural
> >decisions.  Obviously our nomenclatural decisions ARE biased by both
> >geographic distribution *and* historical nomenclature.
>
> I think the first is true (whether it should be or not is another
> matter),
> but I would like to think our science was open minded in terms of prior
> assertions and dispute the second...

Well...I don't know.  If the purpose of scientific nomenclature is optimal
communication, and communication spans generations and history, and
nomenclatural instability disrupts communication.....what is the best
solution?  In my two Centropyge examples above, suppose taxonomic history
treated one set as two species, and another set as one species.  I find that
the patterns seem to be essentially the same in both cases, and my personal
species-boundary-detecting algorithm leaves me right in the middle.  Which
is more important for facilitating communication: preserving nomenclatural
stability maintain historical usage), or imposing some semblance of
consistency (treating both cases in the same nomenclatural way).  I think
most people would go with consistency -- but I still think that preserving
nomenclatural stability has (and perhaps should have) some influence in
making such decisions.

> >Perhaps I have achieved nothing but wasted bandwidth;
>
> RP consume bandwidth!?  Never!

Yeah, but the question is whether I've done anything else in addition to
wasting bandwidth....

> >in which case I
> >sincerely apologize.  But if any of the above sparks interest among list
> >members, perhaps the ensuing discussion might help me understand how to
> >articulate my true question.
>
> Nooooo!   No more with the S word... please!   :)

Sparks? Suede? Shoes? Salamanders?.....Species?

Aloha,
Rich