Re: [asa] historical versus experimental sciences

From: David Campbell <>
Date: Thu Aug 20 2009 - 12:55:37 EDT

> In your examples below, you are granting more knowledge about the species
> (some knowledge of the biochemistry, some knowledge of mitochondrial DNA,
> etc.) than we originally supposed in our scenario.  (Or at least, than I
> supposed.)  All of my argument was based on the assumption that we had *no*
> genetic or physiological knowledge of the new fruit fly, but knew only what
> it looked like, and its eating habits and such behavioural characteristics
> as were obvious to the despairing farmers whose crops were being destroyed.
> My complaint, based on that assumption, was that you would not have enough
> information to determine evolutionary relatives, and your line of diagnosis
> and prescription would thus be stopped in its tracks.  Of course, if you
> have more information, then I understand how you could arrive at
> evolutionary relationships.

If you have whatever information has been adequately studied to be a
pointer to evolutionary relationships, whether it is a DNA sequence or
the configuration of hairs or stripe pattern or mantle flap shape,
then you are able to tell what group it belongs to. [Knowing the
genome means we have the entire thing sequenced, as opposed to having
one or some genes.]

By looking at one feature, we can, on evolutionary grounds, predict
similarities in unrelated features due to common descent. Some
features of practical interest are not as easy to determine directly
as others of less immediate practical interest. For example, if you
capture a few flies in LA, that isn't likely to be enough to
extensively test what attracts them. However, it is easily enough to
generate some DNA sequences. Using the sequence (or possibly the
configuration of hairs or stripes or some other morphological
feature), we can say "aha! these belong to the x group (possibly even
a previously unknown species) and therefore, because they share a
close common ancestor, are likely to be similar in vulnerabilities to
the better-known closely related species."

> However, then the *second* part of my argument would kick in, i.e.:  even if
> you could determine the evolutionary relationships, it would be superfluous
> to do so.  The similarities in DNA and/or and/or general biochemistry would
> point the scientists and policy-makers in the direction of the solution.

Why? If they were created separately, there would be no reason why
the creator could not mix and match physically unrelated attributes to
create a range of forms. There's nothing about similarity in
mitochondrial genes that functionally connects them to what bait
attracts a fly.

It is, of course, empirically true that living things that are similar
in certain ways are often similar in other ways. But if you want to
explain that pattern, evolution is the only model that has worked.
That's why living versus non-living makes a difference.

> I see no benefit in arguing "Often close evolutionary relatives have similar pheromones", if you cannot determine close evolutionary relationship without *first* determining genetic, biochemical or morphological similarity<

Well, determining evolutionary relationship is based on the evidence,
so it is difficult to determine without a consideration of the
evidence. But without evolution you have no explanation for why
similar mitochondrial DNA and similar pheromones (or whatever other
feature you like) go together.

Yes, you can empirically observe that similarity in particular
features often goes together and rely on that without bothering to ask
why the pattern should exist. You can also use Kepler's laws without
ever bothering to formulate a general law of gravity. But the
practice of science is to try to find general common principles that
explain a large number of things. Again, you are making exactly the
same error as the atheists who claim that we shouldn't believe in God
if the laws of science are able to explain science. God provides a
comprehensive explanation of everything, both science and non-science,
and it is no more unreasonable to invoke evolution as explaining the
shared similarities of organisms or to invoke God as behind all the
natural laws than it is to invoke electroweak theory as explaining
both electric/magnetic forces and the weak nuclear force. (Of course,
in each case one must see whether the proposed overarching model works
well, but to dismiss a working overarching model as a mere extraneous
add on is, reductio ad absurdum, to reject all scientific models.)

>But that's exactly what we'd expect if, as Professor Skell has said, as far as most experimental science is concerned, neo-Darwinism is an interpretive gloss which is not necessary to conduct the research or validate the results.<

Any theory, as far as experimental science goes, is not necessary to
conduct an experiment or validate the results. You just go and do the
experiment and check your measurements. The theory tells us what
experiments are likely to be interesting and what the results mean.

> This glossing occurs all the time.  The other day, some scientists
> discovered the gene that gives the water strider its characteristic mode of
> skimming across the surface of the water, and the news story (perhaps
> following the lead of the reporting scientists, I don't know) gave the
> matter an evolutionary twist:  scientists have discovered the evolutionary
> mechanism by which the water strider acquired its ability.  In fact, the
> *empirical* science had uncovered nothing about evolutionary mechanisms at
> all.

They discovered the changes in function in one Hox gene that led to
the shape and length of legs, which enables the water strider to skate
around as it does. In other words, they are providing some of the
information about exactly what genes changed when to produce this
innovation, a piece of information that you have been requesting for
evolution to be validated. Yet now you are saying it is irrelevant
for evolution. (Now it is generally true that the news story probably
didn't have much useful information, but a quick check of
gives the details.)

> I find it interesting that you automatically associate "mitochondrial DNA" with evolutionary theory.<

No, I associate it foremost with "relatively well-studied DNA,
generally informative at the species level," though it is therefore
quite useful for evolutionary analyses (as long as no funny business
is going on, such as hybrid polyploid lineages) .

>  From a non-doctrinaire, non-historical point of view, the mitochondria are simply organelles in the cell which contain DNA,<

Because they evolved from separate free-living organisms

> and the sequences in that DNA can be studied, just as the sequences in
> nuclear DNA can be studied.  One can ask the question "How much different is
> the mitochondrial DNA of *this* species from the mitochondrial DNA of *that*
> species?" without speaking of evolution at all.  It's a question for
> biochemists to settle.

Why should the mtDNA be different? Mitochondria do the same thing in
all organisms that have them; as long as the set of genes work
together properly, there's no need for our mitochondria to be more
like those of apes than mice or even of water striders or land snails.
 They could function if all of them were identical, or if the DNA
similarities were random. But instead there is a regular pattern of
degree of difference, and it correlates to other features that we
would expect to reflect evolutionary heritage. Why should there be
any particular pattern of difference?

> By the way, are you familiar with the work of the evolutionary biologist Paul Hebert at the University of Guelph, who is one of the world's leading experts on genetic bar-coding? < Yes.

> He stated in an interview a few years ago that, to his chagrin, the results of genetic bar-coding using mitochondrial DNA seem to support the view of the "creationists". <

In the context of overselling his project. Ironically, his example of
establishing the provenance of the mouse in the pasta shows that.
What he is saying is that there is little variation within a species
and a lot between species, and he claims that the reason is unclear
evolutionarily but implies that it would fit well with a model
invoking separate creation of species or the like. In fact, there are
good evolutionary reasons to expect the pattern to be common, and
plenty of exceptions to the pattern (with the caveat that pseudogenes
and similar confounding factors that might create the appearance of
greater variation in the barcode are rarely firmly ruled out). Like
myself, he seems to be a morphological taxonomist who got into
molecular work because that was a good source of additional data and a
better way to get funded. As such, he may not have as much background
in population genetics, which provides the evolutionary explanation.

I previously mentioned that one of the things that I ought to be doing
is working on a paper on snails of the genus Juga. As a matter of
fact, the specific bit of the paper that is calling (not loud enough)
right now is to compile a table of the level of "barcode" percent
differences. In these snails, sometimes there is minimal difference
in the barcode (cox1) between species, sometimes there is the good
barcode pattern of slight variation within the species and significant
variation between species, and sometimes there is high variation
between species. The presence of enough variation within house mice
mtDNA to identify where the mouse in the pasta came from shows that
sequences are not, in reality, invariant within species.

Consider a species. There is some degree of variation in the cox1
sequence within it. Divergence within the population eventually leads
to separate species. Because these are finite samples of a finite
population, the original cox1 variation will almost certainly not be
evenly distributed between the two daughter species, especially given
that evolutionarily, there will probably be some degree of correlation
between the general genetic variation in the population and the
distribution of the key feature(s) that promote speciation. Also, the
separation of daughter species often involves some geographic
separation, and the original variation in cox1 probably was not
totally uniform geographically.

As the two daughter species become reproductively isolated from each
other, there is less and less exchange of mitochondrial DNA (although
the occasional hybridization event can throw things off). Within a
reproductively isolated population, alleles will be lost and new ones
will arise due to genetic drift. If the population is small, the loss
or fixation of alleles can happen rapidly (exact rates depend on the
mutation rate, the population size, and the reproductive
pattern-generation time, the proportion of the population that
actually breeds, and the relative contribution of each individual).
If a species has fairly high levels of genetic mixing within it,
fairly low mutation rates, not too large a starting population, and a
long period since splitting off from the ancestor (relative to
generation time), then it ought to have relatively low genetic
variation within the species but medium difference between the

The prime exemplars for barcoding are birds, bats, and insects. But
they are very likely to have relatively long times between speciation
events and good intraspecies mixing, because they can fly and
disperse. If the climate changes at natural rates, they don't have to
speciate to adjust-they can often simply move. Almost none of the
insects known from the Pleistocene glacial periods are extinct today
(prime exception was a specialist mammoth parasite). Birds also are
exceptionally well-studied. They are big, often showy, and and
mostly diurnal. Thus, the chances that we've noticed morphological or
behavioral differences are generally pretty good (except for patterns
only visible with UV).

Additionally, barcoding often gets a bit circular-here are two birds,
currently considered to be the same species, but they have high cox1
differences, so it must be two species according to a barcode paper.
Maybe they are overlooked species, but maybe there's high variation
within the species-you need to go back and see if 1) is the difference
clear and abrupt when you have data for several individuals across the
range, or does it become a continuum? 2) can you find other features
that correlate with the difference?

If you want a genuinely simple evolutionary system, analogous to the
two body problem of launching a satellite, you need something like the
studies that put two kinds of flour beetles into a lab enclosure and
see which one wins out. Depending on precise conditions (especially
moisture), you can indeed predict which one will still be around after
many generations. Predicting what will happen to a lagomorph or a
rodent in the Amazon over millenia is more akin to using the laws of
gravity and motion to predict the exact fate of all the pieces from
the recent satellite-satellite collision.

Dr. David Campbell
425 Scientific Collections
University of Alabama
"I think of my happy condition, surrounded by acres of clams"
To unsubscribe, send a message to with
"unsubscribe asa" (no quotes) as the body of the message.
Received on Thu Aug 20 12:56:41 2009

This archive was generated by hypermail 2.1.8 : Thu Aug 20 2009 - 12:56:42 EDT