Re: [asa] historical versus experimental sciences

From: David Campbell <>
Date: Fri Aug 21 2009 - 11:42:33 EDT

> shows that you do not grasp the redundancy of evolutionary classification
> for the kind of practical matters we are talking about.  The line of
> argument you are using runs this way:

If you take it as a given that organisms that are similar in one way
are likely to be similar in another way, then you can proceed based on
that. If you want to know why that is the case, then you need
evolution. The same is true for any overarching model.

You are correct in saying that one can avoid explicit reference to
evolution while taking advantage of its consequences, but that does
not prove evolution's irrelevance.

> In fact, to be honest, we have no "laws" of evolution at all, but only vague general principles of
> almost infinite explanatory elasticity -- mutation, drift, selection.

In a tightly circumscribed system, they provide quite specific
predictions. The problem is that so many of the obviously interesting
evolutionary systems are complex.

It is true that biological systems in general are less well behaved
than things in physics equations, but that applies across the board.
Medicine, meteorology, and many other fields are no better than
evolution in 100% certain predictions. There are obvious and
generally uninteresting cases where the exact outcome is certain
(e.g., if you put a penguin and a rabbit into the Amazon forest, which
will do better), and then there's general reality.

Conversely, does ID give any useful predictions at all? Individual ID
advocates make specific predictions, but not rooted in the overall
concept of ID, because the latter is too large of a tent to be
specific. Even this does not prove that ID is useless or a mere
add-on, but that merely in and of itself it doesn't give any specific
scientific guidance. There's certainly no guidance as to exactly when
and how any particular intervention would take place.

Evolution can predict what will happen in an intermediately complex
situation. Study of the anolid lizards in the Caribbean shows that
ground-dwelling and tree-dwelling ones on one island are generally
more closely related than either are to those on other islands (with
allowance for plate tectonics, changing sea level, storm transport to
nearby islands, etc.). Take some tree-dwelling anoles and stick them
on a small island with no lizards and few or no trees. Before long
they evolve longer legs and other traits of ground-dwelling anolids.
(Finding islands large enough to have plenty of trees that don't
already have lizards is more difficult.) Analogous situations exist
for arctic fish specializing into different trophic niches, etc.

> You can no more predict an outcome with such vague and general notions as drift, mutation,
> and selection than the scholastics could predict physical outcomes by
> speaking of exemplary causes, formal causes, concurrence, effluents, etc.

As noted above, very specific outcomes can be predicted for a suitably
circumscribed situation. The claim that evolution cannot predict
anything is simply wrong.

Accuracy (right/wrong) versus precision (details) are a problem here.
As Conway Morris has been pointing out a good deal, we can predict
evolutionarily a number of general trends. Fast swimming can be
powered by paddles, by lateral body motion, by vertical body motion,
or by jetting. In any of the cases, the organism needs to be
streamlined. Exactly which organism does what how is more tricky,
although having specific information on the starting features would
enable some prediction (for example, lateral flexing is normal for
early chordates and for most reptiles; vertical flexing is normal for

Suppose that we make Mars habitable, with some ocean and some land.
We stick a small selection of organisms there and monitor over
extended periods of time. While it is probably impossible to predict
exactly what any one organism would do, we can have a pretty good idea
of what some of the general patterns are going to be. Some organisms
would develop fast streamlined swimming; some would go in for
increasing defenses; some would have longer legs for running, etc..

> B.  Regarding the water strider, I took the time to look at the article.  It
> does *not* contain any argument, as far as I can see,
> about "exactly what genes changed when".

They are looking at exactly what genes changed. They did not look
closely at when, but give some measure of it by determining that the
water strider is different from other true bugs. I'd like for them to
specifically compare with the closest relatives that don't have
unusual legs. Nevertheless, this is a part of the information that
you keep asking for.

> Note that, without a record of the genome of the purported ancestors, the
> term "ancestral" is a loaded term, implying the evolutionary relationship
> which the article purports to establish; yet the article provides no data
> about the alleged ancestors.

Paleontological data can give us information about the ancestors, as
can phylogenetic studies on modern forms. For example, I predict,
based on evolutionary studies, that they should next look at one of
the non-strider Gerromorpha to find the most similar "normal" version
of the gene (and not at more distantly related water bugs, nor at
fishing spiders, etc.). They should also look at the Veliidae and
Hermatobatidae to maximize their intra-strider range of variation.
Examination of the differences in sequence between the water striders
and, e.g., the water measurers (which walk but don't skate on the
water and which don't have extra long middle legs) , along with
examination of the interactions with other genes, would pin down the
exact mechanism. The timing is a bit problematic, because the fossil
record of the group is rather patchy, but striders are present before
120 million years ago.

Many insects are lightweight enough relative to the surface area of
their feet that they can walk or run on the surface of the water, and
the close relatives of water striders typically live near or on the
water, so they would already have some component of waterproofing.
The key innovation is being able to skate with the extra long middle
leg. At least some (they're poorly known) of the Hermatobatidae have
a somewhat different pattern of skating, a bit more like normal insect
locomotion, which may preserve some of the transition.

> Were there once water striders that could not skim the water?

Well, the ancestors of water striders once were not able to skate on
the water, though just about any small enough terrestrial arthropod
could walk on water. Whether they ought to be called water striders
at that point is doubtful.

> When did they live?

The non-skimming ancestors of skimming water striders were no later
than 120 million, the oldest fossil (somewhat non-standard) striders,
but probably not more than about 200 million, given the lack of
records from the Triassic rift lakes. If you want more precise dates,
go to something that has a really good fossil record, like mollusks.

There are living groups that probably have similar life style and
appearance to the ancestors of water striders, but I don't think
that's what you are asking about.

> What were their genomes like?

The genomes would be quite similar to a consensus of a water strider
genome and a non-strider gerromorph.

> How do we know any of that?  The article is silent about these things.

What do you expect from molecular biologists? They investigated the
function of a particular gene.

> Rather, it assumes, just because it has located a gene which appears to be responsible for the difference in leg length, that this result should be interpreted in terms of an evolutionary change.  But the evolutionary framework is brought to the data; it does not follow from the data. <

So evolution is providing something additional, not merely from the
immediate data?

> It would only follow from the data if we had reliable genomic and morphological data from the putative ancestors. <

No, we can tell something about evolution from the modern forms,
though there are fossils that provide some morphological detail.
Unfortunately, although there are a few water striders and kin in
amber, amber is no good for DNA preservation. It seals water in, and
dry is what you want to preserve DNA. It's conceivable that an
ancient water strider or its kin might turn up in a salt deposit, but
for now DNA only goes back to the Pleistocene (pending any
verification of claims of ancient DNA in salt deposit bacteria, but
the potential for flow or dissolution and recrystallization in salt
makes excluding contamination extremely difficult). Evidence of
protein or other durable macrobiomolecule sequence is probably
possible in more cases, but as of yet not so well studied, and
probably not often giving resolution at this level.

> C.  Your dismissive remarks about Hebert's motives and training are
> unresearched and clearly ad hoc.  I don't buy them.  Neo-Darwinian theory
> makes a prediction about what we should find in the mitochondrial DNA.
> Hebert, a neo-Darwinian, candidly admits (to his credit) that the data do
> not support the prediction.  Your answer is to refer vaguely to other
> factors which might explain why the Darwinian prediction is falsified.
> That's the problem with neo-Darwinian explanation.  It is too elastic.  It
> can always call upon other vaguely understood possible factors to rescue the
> theory when the data doesn't match.  That is why I maintain that Darwinian
> theory can never be falsified, and is not good science.  A robust, manly
> scientific theory puts its neck on the line, and when it's wrong, admits
> it's wrong.  It doesn't need to always be pleading special excuses, as
> neo-Darwinism does.

Is the problem that evolution doesn't predict anything or that it
makes predictions that are wrong? It can't do both.

If different features showed no correlation between organisms, that
would be good evidence against evolution. Yes, there is an exception
in that bacteria can quite easily, and eukaryotes more rarely, pick up
DNA from other organisms, but even that ought to follow some patterns.

I am doing barcoding-type work on mollusks. I am not merely making up
objections; I know good and well that there are a number of problems
with the barcoding endeavor that Hebert didn't mention. I know that
the claim that the barcode always works is wrong. Of course, this is
science journalism, which as you noted is deficient in caveats. But
the models of population genetics predict the pattern that we observe,
with some specific guidelines as to when we would expect to see what.

In some small fish that Steve Powers was working on, there are clear
morphological difference in the color pattern of the males. It's what
they use to recognize mates, so this gives very good evidence to
consider them separate species. Yet the mitochondrial data do no good
in separating them out. It turns out that the fish have such large
population sizes that there is little chance of an allele totally
disappearing form the population.

Ironically this ties into one of your other objections. The reality
is that evolution is a messy process, with real organisms influenced
by all sorts of parameters, quite unlike a frictionless massless
pulley with a massless rope and a precisely known force. As a result,
some will show one pattern, some will show another.

> Do *anything* but admit that the evidence may be more in favour of intelligent design than of accidental mutations and fortuitous selections.<

Evolution actually provides predictions, unlike ID, and so can be
supported by the evidence. As someone who actually studies fossils,
modern morphology, and DNA, I find that evolutionary models work.

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 Fri Aug 21 11:43:06 2009

This archive was generated by hypermail 2.1.8 : Fri Aug 21 2009 - 11:43:06 EDT