[Taxacom] The 'reality' of species boundaries -- Once Again (UGHHH!)

[Stephen Thorpe]

Richard,
>Where is the analogy in that?
I may have misinterpreted you, but you seemed to be suggesting that because we need to choose a threshold of reproductive isolation (90%, 95%, etc.) from a continuum of possibilities, in order to define species boundaries according to the BSC,  the species boundaries themselves were therefore purely subjective. The analogy was that you do in fact have to choose a significance level for a statistical test from a continuum of possibilities, but the test itself is not thereby rendered purely subjective.

>just explain to me how we determine if there is sufficient hybridization (with as much implied fuzziness and imprecision as you feel necessary) to decide whether we are seeing two separate species with rare hybridization, or one single species with imperfect gene flow
OK, I will: here we go: choose whatever threshold you like (!), but better make it consistent with uncontroversial cases. Having chosen it, the species boundaries are now determined by the world, not by us! We must now look to the world to see what they are. We look for evidence of mechanisms that would result in X % reproductive isolation. If we had chosen differently, then the species boundaries would be different in some cases, but probably very few, as most cases probably have virtually 100% reproductive isolation, which is why chimp and human aren't the same species by any "definition". So there is some leeway, but not much. This is how it is done IN THEORY. IN PRACTICE, we don't decide on a precise threshold and take it from there. Instead, we look for evidence of "very high" levels of reproductive isolation, such as very different genitalic morphology, etc. What is "very high"? That is fuzzy. Nevertheless, we look to the objective world to discover species boundaries in a manner that is totally lacking for genera or other levels. Species are more like Australia, ... Genera are just "convenient monophyletic groups", unless you believe in some sort of well-defined objective notion of "overall similarity", in which case a particular threshold could be chosen and then the genera would also be determined by the world. I don't ...

>> but the very practice of biological taxonomy for sexually reproducing organisms involves trying to discover the natural boundaries of reproductive integrity and calling them species.
>Yes, I know this is how you see it.  But I, and many other taxonomists, see it a different way
I believe that I have just explained in what sense 'the very practice of biological taxonomy for sexually reproducing organisms involves trying to discover the natural boundaries of reproductive integrity and calling them species'  is consistent with your question 'just explain to me how we determine if there is sufficient hybridization (with as much implied fuzziness and imprecision as you feel necessary) to decide whether we are seeing two separate species with rare hybridization, or one single species with imperfect gene flow?'

Regarding hybridisation, and with particular reference to the most recent post by G. Singh, we must be careful to distinguish the case where the hybrid population itself becomes reproductively isolated from the parents, in a new speciation event...

Stephen

________________________________________
From: Richard Pyle [deepreef at bishopmuseum.org]
Sent: Saturday, 12 September 2009 4:32 p.m.
To: Stephen Thorpe
Subject: RE: [Taxacom] The 'reality' of species boundaries -- Once Again (UGHHH!)

> >If that's your answer to the question, then why did you give
> me the bogus answer on statistics the first time around?
> Another "politician-like response" from Dr. Pyle! :) My first
> answer was an analogy, and therefore not "bogus"!

No, it was not an analogy.  Here is what you said:

"so, statistics is all just human subjectivity, is it? Funny, I thought the
whole idea of statistics was to reveal objective facts about the world! But
wait ... how could I have been so stupid? You have to choose a significance
level! So, statistics is all just human subjectivity after all ..."

Where is the analogy in that?  It's a bogus answer because you are
misdirecting (dodging) my specific question.  I asked you what proportion of
hybridization was acceptable in order for there to still be a species
boundary.  You replied with the nonsense quoted above, which I *think* was
an attempt to (snidely) imply that I was saying statistics were subjective
(which clearly this is not what I was suggesting).  It seems to me that you
are the one acting more like a politician here.

> The fact that there isn't a perfectly precise
> objective answer to "which level of significance should we
> use?" in statistics, doesn't make statistics any the less
> objective. See the analogy?

No, I don't see how this is an analogy.  I already addressed the precision
issue in my later response.  Don't give me a precise answer, just explain to
me how we determine if there is sufficient hybridization (with as much
implied fuzziness and imprecision as you feel necessary) to decide whether
we are seeing two separate species with rare hybridization, or one single
species with imperfect gene flow.

> >But...what if the hybrid populations continue to persist?
> >Are they a third species? Or no species at all?
> That wasn't your original question!

Yes, it was!  I asked what would happen if the two allopatric populations
disappeared, and all that was left was the hybrid populations:

"Now, let's say that the broad allopatric populations all die out, leaving
only the hybrid populations, such that 100% of individuals form hybrids.
Are they now suddenly the same species, simply because the allopatric
populations disappeared?"

When I asked "Are they now suddenly the same species" -- the "they" was
referring to the remaining extant individuals in the hybrid zone.  I'm sorry
if this was not clear.

> I'm not sure that there
> is an answer to that, other than just "they are hybrids of
> extinct species".  I would lean to the third species option,
> since some of the species we recognise today could have been
> formed as hybrids with subsequent extinction of parents.

OK, that's a fair answer.  I don't agree with it, but at least it is an
answer to my question.

> Anyway, I think I understand you as arguing that there is a
> complete (or nearly complete) spectrum of interbreeding in
> the world today, so no non-arbitrary place to draw species
> boundaries.

To be precise, I am saying there is a nearly complete spectrum of *examples*
of different levels of interbreeding in the world at all times (not just
today).

> This seems to be based on your own first hand
> explorations in the field. Am I right?

In part -- but just as much (if not more) from my conversations with other
taxonomists working with different groups of organisms, who also have many
years of experience observing nature.

> If so, I can do little
> more than say that my "experience in the field" screams just
> the opposite!

Fair enough.  Perhaps it is simply a function of our different experience.

> I suspect that you have seen a few cases where
> things are less than clear, and this leads you to make the
> assertion that there is a complete continuum!

No, that's not my assertion.  My assertion is that if you look across *all*
groups of sexually-reproducing organisms, you will find *many* examples that
confound a simply BSC view of the world.  If these examples were confined to
just one kind of organism (like fishes), or if they were extremely rare in
nature, I would be more confident that a natural barrier exists for species.
But too many taxonomists in too many different groups have conveyed to me
too many examples of the sort I have provided for you, that it seems to me
that the "fuzziness" of your "Australia" analogy is, on average, much
broader than the intertidal zone and a few islands here and there.  In that
context, your analogy of Australia would be more representative of typical
species if the intertidal zone were hundreds of kilometers wide.

> I
> claim that in something like 95+% cases of sexually
> reproducing organisms, the levels of interbreeding are well
> within the bounds of what is allowable according to a proper
> understanding of the BSC.

I'm assuming that you just made up the "95%" figure -- which is fine; let's
use it.  That means that 5% of cases rely on subjective interpretation by a
taxonomist.  I would say that 1 in 20 species being ambiguous is a
non-trivial proportion of biodiversity.  But I think you're a bit over
confident in your estimation.  From a lumper's view of the world, 95% is
probably about right.  From a splitter's view of the world, I bet it would
be more like 60%.

> That is why we see lions and
> tigers, and why ligers and tigons are rare curiosities.

Yes, but across the spectrum of biodiversity, mammals are more the exception
than the rule.

> It could
> have been different. Yes, there are a few problem cases,
> which you seem to see as negating the general rule,

OK, so at least you concede that you're talking about a "general rule" in
nature -- and on this point we may actually come close to agreement.

> but the
> very practice of biological taxonomy for sexually reproducing
> organisms involves trying to discover the natural boundaries
> of reproductive integrity and calling them species.

Yes, I know this is how you see it.  But I, and many other taxonomists, see
it a different way.  We see evolution as a continuous process, starting some
4 billion years ago and continuing through an utterly unbroken sequence of
reproductive events to the set of extant biodiversity we see today.  We
recognize that the process of "speciation" (i.e., the process by which
lineages of organisms develop reproductive barriers) is highly imperfect,
and riddled with exceptions to any "rule" that biologists have tried to
invent.  We don't see the process of assigning names to organisms as the
discovery of real entities in nature, but rather as a highly effective means
of communication.  In many cases (at least 60%, and perhaps as much as 95%),
there is very little dissention among taxonomists about how best to label
clusters of organisms as "species".  In the remaining 5%-40% of cases, there
is disagreement about how best to label biodiversity.

To be sure, some of this disagreement is due to imperfect knowledge.  With
perfect knowledge, the disagreements would be reduced in number.  But even
with perfect knowledge, evolution is a continuous process in the sense that
every organism that has ever lived is connected by an unbroken sequence of
reproductive events.  If you draw a slice through evolutionary history at
one particular time (like now), then that slice will intersect every single
stage of the evolutionary process.  Your perception is that the proportion
of cases in that slice that are ambiguous with respect to the BSC is small
(5% or less).  My perception -- and the perception of others who have spent
a long time studying the taxonomy of different groups of organisms -- is
that the proportion is not so small.  This, I suspect, leads to our
different perceptions of the degree of "realness" of species.

> Not so
> for genera, etc., which are just "convenient monophyletic
> groups". Interestingly, I don't see Centropyge as being a
> problem case: from what you say, the levels of hybridisation
> are low, with nothing to indicate that the level would
> increase significantly UNDER PRESENT CONDITIONS. Little
> wonder then, that "everyone" considers them to be distinct!

If I had to guess, I would say that something on the order of 10% or so of
all (combined) individuals of both species are hybrids.

Aloha,
Rich