The Non-Randomness of Natural Selection - Does it really matter?
by bipodSo we all know that Richard Dawkins gets a little pissed off when he gets asked whether evolution is completely random
You said in a recent speech that design was not the only alternative to chance. A lot of people think that evolution is all about random chance.
That's ludicrous. That's ridiculous. Mutation is random in the sense that it's not anticipatory of what's needed. Natural selection is anything but random. Natural selection is a guided process, guided not by any higher power, but simply by which genes survive and which genes don't survive. That's a non-random process. The animals that are best at whatever they do"”hunting, flying, fishing, swimming, digging"”whatever the species does, the individuals that are best at it are the ones that pass on the genes. It's because of this non-random process that lions are so good at hunting, antelopes so good at running away from lions, and fish are so good at swimming.
The point that DawkyDawk misses is that natural selection operating on variation is supposed to be an explanation for the historical development of life. Well, there is this tiny little fact associated with the development of life that needs explaining: the emergence of complex functionality.
Mister Dawkins tells us that evolution is "guided not by any higher power, but simply by which genes survive and which genes don't survive. That's a non-random process."
But please note what's going on here: natural selection prunes from randomly generated solutions in an environment where some options fail to propogate. The power of natural selection should be no greater than the power of completely random search in an environment of unlimited resources.
In other words, Darwinism has to be committed to the view that the same features that we see today in the biological world would also be produced in an environment with unliminted resources, where all organisms reproduced and passed on their genes. All the novelty that exists today would exist in that environment, plus some.
That's the intuition behind claims that evolution is random: the generative aspect of evolution is really no better than random search under certain possible environments.
December 14th, 2005 at 3:19 pm
This is why natural selection is not science to me. It simply proclaims that the genes that survive are here because of it. How can you test that? If you run a test and predict that bacteria "˜A' will survive because it is more adaptive to high heat levels, isn't that stating the obvious? It's like saying I predict that non-swimmers will drown when pushed into a lake and swimmers will survive.
It seems to me that natural selection is simply the laws of physics and chemistry in action and not some higher-level "law" that physics and chemistry are subject to.
Comment by Lurker — December 14, 2005 @ 3:19 pm
December 14th, 2005 at 4:07 pm
Theory, i.e., evolutionary theory itself, only attributes, and which is all that is necessary for evolutionary theory, that there are random and there are selected attributions. Theory, theory, does not actually either state or require that the selections that are made, that exist, that we study, are attributable to either "nature" or "design." It can't do that! Theory is indifferent. That's the only way theory is tested.
By design. By testing. Comparing our own artificially selected, our own designs, against what we find to exist in nature. That's Darwin's theory of Natural Selection.
Dawkins is a dunce. He doesn't understand, what he teaches, proselytizes.
When are the IDers going to address the heavyweights? Dawkins is a cartoon character.
When are IDers going to abandon their strict adherence to the anitquated "creationist agenda" and begin to deal positively with the theory of selection?
Comment by Rock — December 14, 2005 @ 4:07 pm
December 14th, 2005 at 5:26 pm
In other words, Darwinism has to be committed to the view that the same features that we see today in the biological world would also be produced in an environment with unliminted resources, where all organisms reproduced and passed on their genes. All the novelty that exists today would exist in that environment, plus some.
Not quite. Your statement might be true in a situation with unlimited resources AND infinite population. Reiterated mutation and selection increases the probability that those features will arise in a finite population in limited time.
That's the intuition behind claims that evolution is random: the generative aspect of evolution is really no better than random search under certain possible environments.
With a sufficiently vague definition of "possible," I suppose. But why should we bother with an intuition about certain "possible" environments when we can ask how selection and random search compare in real environments?
Comment by Nick — December 14, 2005 @ 5:26 pm
December 14th, 2005 at 6:05 pm
Lurker wrote:
I think you are on the right track. Natural selection is not a force at all, it is simply a rhetorical label attached to the results of a process that we do not yet really understand, a process that takes place in the realm of biochemistry and physics (at least insofar as the individual organism is concerned). A label is not bad if it is actually helpful in assisting us to understand the event in question, but "natural selection" rarely confers anything useful to the discussion. Indeed, it generally acts as a "gaps" explanation, by pretending to explain something that has not been explained, only labeled. It thus lulls folks like Dawkins into thinking that they have an explanation, when in fact they have only an illusion of one.
Ultimately, it doesn't matter how cute we try to be in trying to talk about differential reproduction, tendencies, phenotypes, population averages, etc. — in almost all cases we are still left, as was Darwin, trying to explain the "fit" in terms of those that survive, which of course means nothing at all.
Comment by Eric Anderson — December 14, 2005 @ 6:05 pm
December 14th, 2005 at 6:31 pm
Eric Anderson,
Thanks for the comments. I thought maybe I was the only one who saw it this way. Adding a label to some process and then pretending that you've learned more about that process is ridiculous.
Natural selection just tells me what I already know - that weak life will die when required to be strong, that fragile life will die when required to be tough, that antibiotic resistant life will survive when confronted with the antibiotic.
Comment by Lurker — December 14, 2005 @ 6:31 pm
December 14th, 2005 at 6:43 pm
Nick,
Yes, you are right on all accounts.
Comment by bipod — December 14, 2005 @ 6:43 pm
December 14th, 2005 at 7:09 pm
Natural selection doesn't do anything original or creative. All it does is throw stuff out, so it is very misleading to suggest that it is in any way a "non-random creative force" as some would have us believe.
The best natural selection can do is keep things from getting worse (i.e., prevent devolution), by throwing out the bad effects of mutations before they have a chance to propagate. This explains why we find blind fish with atrophied eyes in dark sea caves. (There is no disadvantage to being blind if there is no light with which to see.) Genetic errors that caused the blindness were not selected against.
Thus, a lack of selection can explain why fish have eyes that don't work, but the presence of selection can't explain where vision came from in the first place.
Comment by GilDodgen — December 14, 2005 @ 7:09 pm
December 14th, 2005 at 7:48 pm
I agree with the above remarks. Natural Selection is not a non-random creative force, it's a non-random sustaining one.
Comment by Dane Parker — December 14, 2005 @ 7:48 pm
December 14th, 2005 at 11:14 pm
Why do you claim original innovations, like T-Urf13 or the genes that confer DDT resistance to mosquitoes, are not creative?
Beating DDT was no mean trick.
Comment by edarrell — December 14, 2005 @ 11:14 pm
December 15th, 2005 at 9:05 am
Dawkins sez:
Natural selection is a guided process, guided not by any higher power, but simply by which genes survive and which genes don't survive.
Which is funny because the same guy tells us that NS is blind and purposeless:
(from "The Blind Watchmaker")
How many people know guides who are blind and without purpose?
NS didn't create Turf13 or the genes that confer DDT restistance. It just allowed those traits to stay around by selecting the organisms that had/ have them (that is if those traits offered an adavantage in that populations' enviro)
It should be noted that before Darwin scientists saw NS as a conserving factor (Ed Blythe).
Then there is the measurement factor- how can we measure NS? Well it looks like some scientists have tried:
The Strength of Natural Selection in the Wild
Comment by Joe G — December 15, 2005 @ 9:05 am
December 15th, 2005 at 10:48 am
GilDodgen:
The best natural selection can do is keep things from getting worse (i.e., prevent devolution), by throwing out the bad effects of mutations before they have a chance to propagate.
If all mutations were deleterious, your conclusion about natural selection would be correct. Based on what we know of molecular genetics, I do not believe that it is even a theoretical possibility that all mutations are deleterious. Perhaps you could elaborate. Since both neutral and beneficial mutations have been observed, your conclusion about NS is false; natural selection can also increase the frequency of some novel genes and alleles (as JoeG points out).
JoeG:
NS didn't create Turf13 or the genes that confer DDT restistance. It just allowed those traits to stay around by selecting the organisms that had/ have them (that is if those traits offered an adavantage in that populations' enviro)
Correct. Therefore, the combination of mutation and natural selection alters the genetic structure of a population. If this thread continues, we'll end up with a description of NS that differs not at all from what Dawkins says.
Comment by Nick — December 15, 2005 @ 10:48 am
December 15th, 2005 at 11:04 am
Nick,
Do you agree with these statements (these are not rhetorical…I'm just feeling out where to take this):
1. If every organism were able to reproduce at close to equal rates and the population grew exponentially, then the genetic structure of the population would not have the same capacity to produce the variation we see under natural selection
2. In order for the natural selection to be rendered impotent (on this line of reasoning), we'd have to have some conclusive way of knowing whether or not, given infinite resources and an infinite population, random search would be be able to produce the current set of biological systems plus some.
3. That some things are nomologically impossible on mere random variation given the laws of the our universe (some things would just never occur even with infinite resources)
That's a start. Could you comment?
Comment by bipod — December 15, 2005 @ 11:04 am
December 15th, 2005 at 11:40 am
This is what I was taught about NS:
However some selection process also acts on non-heritable traits.
Then there is the "survival" question- that being, surival of the individual or the population or the species? Survival of the individual would be the easiest to explain, however survival or the population or species requires knowledge of such things. Knowledge doesn't arise via mutation…
Comment by Joe G — December 15, 2005 @ 11:40 am
December 15th, 2005 at 3:05 pm
Do you agree with these statements (these are not rhetorical"¦I'm just feeling out where to take this):
Population genetics isn't really my thing, but I'll give it a shot.
1. If every organism were able to reproduce at close to equal rates and the population grew exponentially, then the genetic structure of the population would not have the same capacity to produce the variation we see under natural selection
If every organism actually did reproduce at close to equal rates (not just was able to) and the population increased exponentially, then presumably the main influences on the genetic structure of the population would be new mutation and genetic drift. The overall frequency of any new allele in the population would remain low relative to alleles that were present in earlier generations. I'm not sure what exponential growth would do to genetic drift, but I guess it would depend on the size of the founder population. If the founder population was small, you would presumably see loss of some alleles present in the first generation and fixation of others, even if every animal reproduced equally.
So yes: the population that you described would not look much like a population with differential reproductive success. The closest real world scenario that I can think of would be a population that is completely inbred with not genetic variation whatsoever. In that situation, natural selection would be impotent until new mutations generated variation.
2. In order for the natural selection to be rendered impotent (on this line of reasoning), we'd have to have some conclusive way of knowing whether or not, given infinite resources and an infinite population, random search would be be able to produce the current set of biological systems plus some.
I'm not exactly sure what this line of reasoning is, since "2" doesn't seem to follow from "1." But, if you are saying that it is not possible to generate a critique of natural selection from this sort of hypothetical scenario, then I agree, because infinite populations with infinite resources are so completely unlike real world populations. A critique of natural selection, if it isn't drawn from real examples, should at least model real populations.
3. That some things are nomologically impossible on mere random variation given the laws of the our universe (some things would just never occur even with infinite resources)
I'm not sure what to say about this, since I think that infinite resources are probably impossible in our universe. If truly infinite resources are impossible, then I don't think it is meaningful to discuss what would or would not be possible with infinite resources.
However, I suppose we could discuss the idea that genomes are finite, so that in principle, there is a finite (though very large) number of individual steps between any two genomes. We might use SELEX as an artificial model where reiterated selection makes very probable an outcome which is highly improbable given a random search.
Comment by Nick — December 15, 2005 @ 3:05 pm
December 15th, 2005 at 3:16 pm
Bipod says Darwinism has to be committed to the view that the same features that we see today in the biological world would also be produced in an environment with unliminted resources, where all organisms reproduced and passed on their genes.
Your hypothetical world is impossible, because it would require that even organisms which lack the genes to reproduce would reproduce anyway. Your hypothetical could only work if you engage in selection by prescreening out organisms lacking reproductive capacity. You would also have to preselect against organisms which are so defective that they die before reaching sexual maturity. You would also have to preselect against organisms which consume other organisms for food, to prevent organisms being killed before they can reproduce. In a population of replicating organisms, natural selection can only be eliminated by artificial selection or a population of perfect replicators which are all exact copies of each other, in which case there would only be one kind of organism. Since your hypothetical is impossible, Dawkins doesn't have to concede anything.
Comment by Aagcobb — December 15, 2005 @ 3:16 pm
December 15th, 2005 at 4:36 pm
aagcobb.
are you saying that perfect replicators could not have random variation?
Comment by bipod — December 15, 2005 @ 4:36 pm
December 15th, 2005 at 4:43 pm
Yes, bipod. If they replicate perfectly, there is no variation; they are just clones of each other.
Comment by Aagcobb — December 15, 2005 @ 4:43 pm
December 15th, 2005 at 4:52 pm
I'm just trying to get a feel for the effect that natural selection has on the capacities of random variation to produce certain content.
NS prunes the number of replicators. It also tends to homogenize the population. In doing these things, it modifies the genetic structure of the population.
Here's a question: if natural selection played a weaker role (by this I just mean to consider a scenario in which a population is growing, without predation, and with abundant resources) what would the effects be on a population regarding the development of novel functionality, even if that functionality were to only emerge *once* and not be spread across the population.
Should we expect more or less of this "one timer" emergence of novel functionality on a population that is largely immune to natural selection?
Comment by bipod — December 15, 2005 @ 4:52 pm
December 15th, 2005 at 4:54 pm
Ok. Sorry, I hadn't understood what you meant be perfect replicators.
Comment by bipod — December 15, 2005 @ 4:54 pm
December 15th, 2005 at 5:02 pm
To clarify my response to Bipods proposition 3:
My understanding of "nomologically impossible" is that it refers to things that are empirically impossible in the real universe. Since infinite resources seem to me to be nomologically impossible, I can't really say anything about what is or is not nomologically impossible given infinite resources.
I would agree that given finite resources and a finite population, some things are so incredibly improbable as to be nomologically impossible on random variation alone. But, this doesn't seem relevant to the question of whether random variation plus natural selection can account for the organisms that we actually observe.
In a hypothetical scenario of infinite population and infinite resources, we might talk about logical impossibility rather than nomological impossibility. In a scenario where all organisms breed equally (thus no genome variant is deleterious) and resources are infinite, the question boils down to whether it is logically impossible for an infinite search space to contain a finite number of genome variants. I do not think this is a logical impossibility, but I don't see how that is relevant to natural selection either. The hypothetical situation remains the same whether real organisms evolve, were intelligently designed, or were created ex nihilo last wednesday.
And now my brain hurts, so I'll stop.
Comment by Nick — December 15, 2005 @ 5:02 pm
December 15th, 2005 at 5:07 pm
Your hypothetical world is impossible, because it would require that even organisms which lack the genes to reproduce would reproduce anyway.
Then modify the world so that natural selection only serves to weed out organisms which lack the genes to reproduce. I don't think this effects the thought experiment at all. In fact, all I'm looking for at this point is to identify differences between scenarios in which natural selection is active in a population and scenarios in which natural selection is largely inactive.
Your hypothetical could only work if you engage in selection by prescreening out organisms lacking reproductive capacity.
Ok. Then let's do that. So, instead of eliminating natural selection, let's weaken its role a bit and consider the effects.
You would also have to preselect against organisms which are so defective that they die before reaching sexual maturity.
I'll give natural selection that role too.
You would also have to preselect against organisms which consume other organisms for food, to prevent organisms being killed before they can reproduce.
I don't think this is true, if what we're trying to get at is the effect of the absence or presence of natural selection on random variation. It is sufficient to have just one population that sees unrestrained, exponential growth.
Comment by bipod — December 15, 2005 @ 5:07 pm
December 15th, 2005 at 5:26 pm
Actually, Nick, I may still want to hold on to nomological possibility only because I'm interested in possibility under the constraints of our universe's physical conditions. So let's throw out infinite resources and turn it into abundant resources. Yeah, I think that's the right thing to do. Go for minimalism about the requirements for the thought experiment.
What I'm really after, of course, is getting a grasp on what we might expect to emerge from a population of organisms that is largely immune from natural selection. Would such a population, growing exponentially without strong selective pressures, eventually produce lots of new functionality that may only be had by a few members of the population but still be had (because of more opportunities for variation)?
My head hurts too and I know I've made a mistake.
But I think I can sum up my concerns in one paragraph: for a thing to "guide" it must be able to provide some sort of real causal force. Mutuations and other means of variation seem like real things operating at the level of physics and chemistry. They've got real causal force. It doesn't seem clear to me how the "chain of dominoes" gets modified by natural selection.
Comment by bipod — December 15, 2005 @ 5:26 pm
December 16th, 2005 at 7:53 am
bipod says You would also have to preselect against organisms which consume other organisms for food, to prevent organisms being killed before they can reproduce.
I don't think this is true, if what we're trying to get at is the effect of the absence or presence of natural selection on random variation. It is sufficient to have just one population that sees unrestrained, exponential growth.
It is true, bipod. A huge part of natural selection causing populations to diversify involves the arms race between predator and prey. One of the biggest elements in an organisms ecosystem is other organisms they are trying to eat or which are trying to eat them. If you don't preselect against predation, then natural selection is going to play a huge role in how organisms evolve in your hypothetical world. Which means random variation without selection is not going to produce as much diversity, there will be less, because your organisms will be limited to those which get their nutrition from the sun like plants or those which get it from chemical reactions; as soon as predation appears in your world, natural selection will start playing a huge role.
Comment by Aagcobb — December 16, 2005 @ 7:53 am
December 16th, 2005 at 11:13 am
Bipod,
What I'm really after, of course, is getting a grasp on what we might expect to emerge from a population of organisms that is largely immune from natural selection.
If I understand correctly, you are describing a population where mutation and genetic drift are the only influences (population A), and you want to compare that to a population under selective pressure (population B). Relative to population B, I think population A would maintain higher genetic diversity (at least in those genes relevant to the selective pressure in population B). In population B, natural selection would be constantly reducing the frequency of some alleles and increasing the frequency of other alleles. Genetic drift would tend to do the same thing in population A, but less efficiently.
With regard to evolution of novel functionality, which population would generate more would probably depend on the size of the two populations and the number of independent mutational events involved in generating the new functionality. If populations A and B are the same size, then the frequency of single events should be roughly equal. However, since favorable mutations will spread more rapidly in population B, subsequent mutations are more likely to occur in the modified background. Population A would require a much larger base population for multiple mutations to be present in the same background, but in nature, most animal populations are relatively small.
If we were to split each population into sub-populations (A1 A2 B1 B2) and subject B1 and B2 to different selective pressures, then I think we would observe B1 and B2 diverging much more rapidly than populations A1 and A2.
Incidentally, there was a nice paper in Nature recently describing teterodotoxin resistance in garter snakes — Geffeney et al, 2005, Nature 434:759-763. In this case, the novel functionality is the ability to eat poisonous newts. Tetrodotoxin resistance is mediated by a series of sodium channel mutations, and the more mutations that are present in a single sodium channel gene, the higher the resistance to tetrodotoxin. I would be willing to bet that if you looked at a large enough sample of garter snakes from areas that lack toxic newts, you would occasionally find the same changes and probably some others, due to random mutation. But, the combination of multiple mutations in a single gene is likely only with non-random natural selection resulting from the presence of toxic newts in the snakes' hunting range.
That's just off the top of my head. If you continue to be interested in these issues, I think what you really want is a textbook on population genetics. We are unlikely to come to any resolution by just batting the problem back and forth on the internet.
Comment by Nick — December 16, 2005 @ 11:13 am
December 16th, 2005 at 11:15 am
The following is a good article on NS:
The Strength of Natural Selection in the Wild
The point being is that any heritable advantage is more likely to be lost in a population than it has of becoming fixed. The larger the population the better the chances of this happening (beneficial traits getting lost).
It should also be noted that "beneficial" is a relative word. IOW what is beneficial to an individual in one population in one enviro wouldn't be so to another individual in another enviro.
Comment by Joe G — December 16, 2005 @ 11:15 am
December 16th, 2005 at 11:21 am
Sorry to follow up to myself…
Further to the garter snake example, I predict that if you took another species of garter snake that is normally tetrodotoxin sensitive and subjected it to the same selective pressure (toxic newts), you would again see evolution of tetrodotoxin resistance, although the specific sodium channel mutations might not be identical. That non-randomness of natural selection is why Dawkins objects to claims that evolution is wholly random.
(The snakes in the Nature paper were Thamnophis sirtalis. Some populations of another species, Thamnophis couchii, whose range overlaps with toxic newts are also tetrodotoxin resistant).
Comment by Nick — December 16, 2005 @ 11:21 am
December 16th, 2005 at 11:47 am
Nick:
Further to the garter snake example, I predict that if you took another species of garter snake that is normally tetrodotoxin sensitive and subjected it to the same selective pressure (toxic newts), you would again see evolution of tetrodotoxin resistance, although the specific sodium channel mutations might not be identical.
Unless of course all the snakes in the population die from consuming the newts (before the tetrodotoxin resistance took hold). That would halt the process pretty quick.
I would predict that unless the mutation(s) required to convey tetrodotoxin resistance are directed via some "built-in response to environmental cues" (ala Dr. Spetner in "Not By Chance") that the population would perish before tetrodotoxin resistance arose (in that population).
Comment by Joe G — December 16, 2005 @ 11:47 am
December 16th, 2005 at 12:35 pm
I would predict that unless the mutation(s) required to convey tetrodotoxin resistance are directed via some "built-in response to environmental cues" (ala Dr. Spetner in "Not By Chance") that the population would perish before tetrodotoxin resistance arose (in that population).
Your prediction is proven false for a variety of reasons. Garter snakes eat a wide variety of small vertebrates and invertebrates, so no population is wholly dependent on newts. In some situations, tetrodotoxin-sensitive snakes co-exist with toxic newts, but they are unable to exploit the potential prey source that the newts represent. Tetrodotoxin poisoning may temporarily paralyze a sensitive snake, increasing its risk of predation but not resulting in 100% lethality. Snakes seem to be able to detect the toxicity of the newts, so sensitive snakes can reject toxic newts; it seems that most exposure to tetrodotoxin from the newt skin occurs during swallowing, not in the stomach, so snakes that sense the numbing effect of the newt may be more likely to regurgitate the prey.
So, we know that the presence of toxic newts will not automatically exterminate the garter snakes. We know that sodium channel genes are subject to mutation in all organisms that have been examined to date. We know that specific point mutations will render snakes resistant to tetrodotoxin. We know that for resistant snakes, the newts will represent a rich source of prey that cannot be not exploited by many other predators, including their TTX-sensitive conspecifics. Oh, we also know that snakes which prey on toxic newts become toxic themselves and probably receive some protection from predators.
So which is a more likely scenario?
1. That in populations of snakes, occasional random mutations generate tetrodotoxin resistance. Those mutations are advantageous in populations that co-exist with newts, and therefore increase in frequency. As a result, we tend to observe resistant snakes in areas with toxic newts.
2. Sodium channel mutations occur by some unknown and unprecedented mechanism in which the snake genome somehow senses the presence of toxic newts and makes specific changes to a neuronal gene.
In scenario two, we should always observe resistant snakes when we see toxic newts, because the "built-in" response will always be present. In scenario 1, we should sometimes see resistant snakes in the presence of toxic newts, because of the random nature of the mutations. Natural selection cannot operate if, by chance, the mutation does not occur.
Hmm, which scenario looks more like the real world?
Comment by Nick — December 16, 2005 @ 12:35 pm
December 16th, 2005 at 1:34 pm
Nick,
Thanks. I find your comments very insightful and edifying.
Comment by bipod — December 16, 2005 @ 1:34 pm
December 16th, 2005 at 1:54 pm
The snake/newt example is interesting. Unfortunately, there isn't enough evidence to support that such is the kind of evolution that can lead to 'macro-evolution', since the snakes that are more immune to the neuro-toxin move more slowly than the ones that are not. Obviously — not an unconditional leap in advantage.
Additionally, it would help to know what mutation triggers this resistance. Is it actually a mutation that causes a reduction in relevant neuro-transmitters, for instance? If so, any mutation that causes a reduction in activity or production of anything to cause an "advantage" is only an 'advantage' in the sense that such is gained as a result of 'damage' to the genetic information. Not how you evolve new species…
Comment by Dane Parker — December 16, 2005 @ 1:54 pm
December 16th, 2005 at 2:06 pm
Joe G:
The following is a good article on NS:
Probably better to look at the original article, rather than Berlinski's gloss.
http://kingfish.coastal.edu/bi...
First of all, the paper looked at a specific subset of selection studies. "These criteria exclude many important and interesting studies of selection." Second, the paper found that the majority of published studies lacked statistical significance (see figure 3). If you look at figure 3, you do indeed see an exponential curve with the vast majority of studies showing low selection strength. But, if you limit the analysis to those showing statistical significance (the black curve), the curve becomes much more flat.
It also seems to me that the specific studies that were sampled (wild populations with continuous variation within the population) would tend to be biased towards traits under low levels of selection. If strong selection rapidly fixes a trait in the population, then you will be much less likely to find wild populations that are polymorphic for such traits, even if traits under strong selection arise fairly frequently.
You could set up laboratory experiments to study strong selection by combining two populations that are fixed for different variants of a trait, in order to make an artificial hybrid population. But, those types of studies were excluded by the Kingsolver paper.
So, I think that based on the Kingsolver paper, one can conclude that the studies to analyze selection strength are frequently not significant. I'm not sure you can conclude that selection itself is usually not significant.
It should also be noted that "beneficial" is a relative word. IOW what is beneficial to an individual in one population in one enviro wouldn't be so to another individual in another enviro.
Oh, absolutely. In the case of the garter snakes, there is some evidence that TTX resistant snakes are also slower. So, in an environment with toxic newts, the mutation would be advantageous, but in an environment lacking toxic newts the mutation might be neutral or deleterious. The different selective pressures would tend to magnify the differences between the populations.
A further complication to the whole situation is that predation by garter snakes could drive selection for stronger toxicity in the newts, which would in turn drive selection for greater resistance in the snakes. Co-evolution.
Comment by Nick — December 16, 2005 @ 2:06 pm
December 16th, 2005 at 2:26 pm
Obviously "” not an unconditional leap in advantage.
As Joe points out, NO trait is unconditionally advantageous. Advantage is ALWAYS relative to environment.
Additionally, it would help to know what mutation triggers this resistance. Is it actually a mutation that causes a reduction in relevant neuro-transmitters, for instance? If so, any mutation that causes a reduction in activity or production of anything to cause an "advantage" is only an "˜advantage' in the sense that such is gained as a result of "˜damage' to the genetic information. Not how you evolve new species"¦
It is a series of point mutations in a voltage-gated sodium channel, not a neurotransmitter The specific amino acids are listed in the Nature paper. There is no reduction in protein levels; the amino acid changes prevent tetrodotoxin from binding to and blocking the channel pore. Similar mutations occur in other tetrodotoxin resistant organisms. Some TTX-resistance mutations do make slight changes to channel kinetics, but the protein still functions as a sodium channel. Bear in mind that only some vertebrate sodium channels are TTX sensitive (e.g. skeletal muscle and some neuronal channels). Others are naturally TTX resistant (e.g. cardiac and some neuronal sodium channels). So you can only characterize TTX resistance as a "damage" to the genetic material if you have in mind some idealized sodium channel in an idealized environment. In the real world, everything is relative.
FWIW, the channel in question is the garter snake skeletal muscle sodium channel. If I wrote neuronal earlier I apologize — a slip based on the fact that both muscle and neurons contain TTX sensitive channels.
With regard to evolution of new species, speciation is related to changes in chromosome number, breeding behavior, etc — things that induce reproductive isolation. It is not directly linked to the evolution of new traits. However, once speciation has occurred, new traits that occur in one species cannot transfer to the other. Thus differences between the two species will accumulate. Since the discussion involved the effects of natural selection or random mutation within a population, your comment on speciation is something of a red herring.
Although the Nature paper that I originally referred to is only available with a subscription, several other papers seem to be available for free, and most of my comments are drawn from them. Check them out. There's some really cool data:
Fun fact: the authors include Edmund D Brodie, JR and Edmund D. Brodie III. A father and son science team!
http://www.bio.indiana.edu/~br...
http://www.harding.edu/USER/pl...
http://www.bio.indiana.edu/~br...
Comment by Nick — December 16, 2005 @ 2:26 pm
December 16th, 2005 at 2:48 pm
OK, one last paper and then I'll shut up. Dane asked whether the mutations affecting TTX resistance damaged other sodium channel function. The kinetics of the TTX resistant channels in the snakes do not seem to have been tested, but here is a relevant paper from an analogous scenario:
Bricelj et al (2005) Nature 434: 763-672
A single point mutation makes clams resistant to saxitoxin. The kinetics of the mutant channel were tested, and no changes could be detected. The mutation is advantageous to clams in regions with toxic plankton. So, in this case, we seem to have an advantageous mutation (gain of STX resistance) that is not accompanied by any "damage" to another trait.
Comment by Nick — December 16, 2005 @ 2:48 pm
December 16th, 2005 at 3:27 pm
I would predict that unless the mutation(s) required to convey tetrodotoxin resistance are directed via some "built-in response to environmental cues" (ala Dr. Spetner in "Not By Chance") that the population would perish before tetrodotoxin resistance arose (in that population).
Nick:
Your prediction is proven false for a variety of reasons. Garter snakes eat a wide variety of small vertebrates and invertebrates, so no population is wholly dependent on newts.
In that case there isn't any "selective pressure" due to the toxic newts. That was the point (your point) I was addressing- toxic newts being a "selective pressure". They would be only if they are the sole source of food, less so if they are one of very few sources and even less if there is a wide variety of foos sources.
Nick:
1. That in populations of snakes, occasional random mutations generate tetrodotoxin resistance. Those mutations are advantageous in populations that co-exist with newts, and therefore increase in frequency. As a result, we tend to observe resistant snakes in areas with toxic newts.
2. Sodium channel mutations occur by some unknown and unprecedented mechanism in which the snake genome somehow senses the presence of toxic newts and makes specific changes to a neuronal gene.
In scenario two, we should always observe resistant snakes when we see toxic newts, because the "built-in" response will always be present.
Not so. Ya see that "built-in" mechanism could be messed up in some individuals due to random mutations. Also even if the mechanism was intact it may not have been triggered in all individuals for any number of reasons- one being not every offspring inherits every trait that each parent has.
Comment by Joe G — December 16, 2005 @ 3:27 pm
December 16th, 2005 at 3:30 pm
Interesting Nick. I do not doubt that advantageous occur. And, as I often say, I do not necessarily have a problem with evolution as an explanation for the diversity of life. I am highly skepical of the Darwinian mechanism, and am therefore likewise skeptical that such a mechanism causes anything that one calls evolution (i.e. the process that lead to the lifeforms we now see).
So, when you say that we're looking at a mutation that changes the sodium channel that prevents TTX from binding, that's not a creative process. There is no new function, but rather is simply a diversification of of an already-in-place system that understandably affects related process, such as the binding of TTX for one. You ourself seem to understand the implication here when you admit, "So you can only characterize TTX resistance as a "damage" to the genetic material if you have in mind some idealized sodium channel in an idealized environment".
Now, I'm not sure that I would call it the result of 'damage' to genetic instructions at this point — perhaps it is just the slight diversification of it? But toxin resistance at a price of slower movement as a result of mutation is one thing, while a mutation resulting in the evolution of new genetic instructions coding for an enzyme that breaks down TTX with the snake maintaining its speed is another.
Comment by Dane Parker — December 16, 2005 @ 3:30 pm
December 16th, 2005 at 4:22 pm
They would be only if they are the sole source of food, less so if they are one of very few sources and even less if there is a wide variety of foos sources.
Not necessarily so. In addition to the variety of food sources you also have to consider effective population sizes of those food sources and potential competitors for those food sources. If the newts are toxic to most vertebrates, their population may be quite high relative to other potential snake food. I have seen very dense populations of eastern newts, a related toxic species, in some lakes. So, a resistant snake will have an increased source of food relative to the sensitive conspecifics with which it competes. A sensitive snake will waste relatively more time and energy pursuing prey that it will need to regurgitate and which could paralyze or possibly kill it. If all snakes are sensitive, they are at the same relative disadvantage, but a resistant mutant is not.
And don't forget that the resistant snake has a second potential advantage. It becomes toxic itself if it eats toxic newts.
Not so. Ya see that "built-in" mechanism could be messed up in some individuals due to random mutations.
Are you serious? So, now you have targeted mutations of a sodium channel driven by an unknown "built in" mechanism which is itself affected by random mutations. The alternative is random mutations in the sodium channel which actually mediates the phenotype in question. William of Ockam, paging William of Ockam
Your "built in mechanism" only works if there is actually an advantage to the snakes which are TTX resistant. If you are correct that there is no advantage to resistant snakes, then there is no reason for a "built in" mechanism to generate resistant snakes. If there is an advantage to resistant snakes, then there is no need to an additional "built in" system.
Tell me more about the built in system. How does it work? What is it made of? How does it know which mutations to make in an organism exposed to tetrodotoxin versus one exposed to saxitoxin or pyrethroids? How does it know to mutate a sodium channel gene instead of a calcium channel gene?
Comment by Nick — December 16, 2005 @ 4:22 pm
December 17th, 2005 at 9:41 am
They would be only if they are the sole source of food, less so if they are one of very few sources and even less if there is a wide variety of food sources.
Nick:
Not necessarily so. In addition to the variety of food sources you also have to consider effective population sizes of those food sources and potential competitors for those food sources. If the newts are toxic to most vertebrates, their population may be quite high relative to other potential snake food. I have seen very dense populations of eastern newts, a related toxic species, in some lakes. So, a resistant snake will have an increased source of food relative to the sensitive conspecifics with which it competes. A sensitive snake will waste relatively more time and energy pursuing prey that it will need to regurgitate and which could paralyze or possibly kill it. If all snakes are sensitive, they are at the same relative disadvantage, but a resistant mutant is not.
You just confirmed what I said. The "selective pressure" is directly related to the population size of the toxic newts in relation to all the other food sources in the area.
As for how "built-in responses to environmental cues" would work- very similar to computer programming (C++ for example) with its statements such as "if", "else", "while", etc.
As for Occam what are the odds a "random" mutation would come along in just the right locus (loci) to convey an advantage? What are the odds if several mutations are required?
The most likely scenario is that the adavantage already existed in the population but that advantage became more prevelant when those without the advantage started to be selected out. How the advantage arose isn't important just that an advantage existed and NS acted on it.
Comment by Joe G — December 17, 2005 @ 9:41 am
December 17th, 2005 at 10:21 am
You just confirmed what I said. The "selective pressure" is directly related to the population size of the toxic newts in relation to all the other food sources in the area.
You claimed that the toxic newts were only a selective pressure if they were the sole food source. But, I agree with your new statement. If the newts are so rare as to be a vanishingly small potential food source, then there would be little or no selection for resistant snakes. How does this help your argument that natural selection does not explain TTX-resistant snakes?
As for how "built-in responses to environmental cues" would work- very similar to computer programming (C++ for example) with its statements such as "if", "else", "while", etc.
I've been giving you examples of real biological systems and arguing in terms how how the animals interact with specific environmental features. You're giving me a vague analogy to computer programming? How, in terms of cellular structures and biochemistry, does "if," "else," and "while" allow a cell to know that making targeted mutations of a sodium channel (as opposed to calcium channels or any other protein) is advantageous in the presence of toxic newts?
As for Occam what are the odds a "random" mutation would come along in just the right locus (loci) to convey an advantage? What are the odds if several mutations are required?
That would depend on the population size, the mutation rate, and the number of generations involved. These are single-nucleotide changes that even the most hard-core young earth creationist agrees are possible. As you point out, the mutation could already be present at low levels in the population before that came in contact with the newts — I have already suggested that these mutations probably occur as rare events in all populations of garter snakes (and in all vertebrates, for that matter). As for the second and subsequent mutations, their probability would be the same as the first mutation, because they would occur in a context where the first mutation has already been selected for.
Also remember that there are multiple potential mutations, not just one, that can render a channel TTX resistant.
The most likely scenario is that the adavantage already existed in the population but that advantage became more prevelant when those without the advantage started to be selected out. How the advantage arose isn't important just that an advantage existed and NS acted on it.
So, we're agreed that Natural Selection could result in increased prevalence of the trait in the population? This takes us back to the beginning of the thread. What exactly is your objection?
But where did that TTX-resistant channel come from originally? Your vague built-in mechanism won't explain it, if it is already present in the population before they encounter toxic newts. Mutation happens — has any other real alternative been proposed?
Comment by Nick — December 17, 2005 @ 10:21 am
December 17th, 2005 at 2:14 pm
Nick wrote:
"That non-randomness of natural selection is why Dawkins objects to claims that evolution is wholly random."
I apologize that I haven't followed the entire thread closely, but the "non-randomness" of evolution is only trivially due to natural selection. We know that evolvability is a quality of variation, so the non-randomness in evolution is due primarily to non-randomness in variation. We know this through empiricism. The Darwinian algorithm is often impotent. This is known as the representation problem in evolutionary computation. The solution to the representation problem was the creation of the genotype-to-phenotype map. The map itself creates evolvability, and different maps create evolvability with different trajectories. This is all independent of Natural Selection; Natural Selection does NOT shape potential evolutionary trajectory; it actualizes existing potential. So yes, in this sense NS provides directionality (non-randomness), since nature selects the roads to travel. But the roads were paved by evolvability, which again, is a quality that absolutely depends upon variation. This is why bibod is correct when he says upfront:
"In other words, Darwinism has to be committed to the view that the same features that we see today in the biological world would also be produced in an environment with unliminted resources, where all organisms reproduced and passed on their genes. All the novelty that exists today would exist in that environment, plus some."
Selection streamlines evolution, but it is not responsible for it. The nature of variation creates evolvability. And yes, in a thought-experiment world with unlimited resources, and WITHOUT selection . . . evolution still occurs. So to attribute evolution to natural selection is a false attribution.
Comment by jazzraptor — December 17, 2005 @ 2:14 pm
December 18th, 2005 at 9:16 am
Nick:
So, we're agreed that Natural Selection could result in increased prevalence of the trait in the population? This takes us back to the beginning of the thread. What exactly is your objection?
I'm not sure I had an "objection" to NS. My first post was about Sir Richard claiming that NS was a "guiding force"- that was from the same guy who told us that NS is blind and purposeless.
How many guides do you know who are blind and without purpose?
Comment by Joe G — December 18, 2005 @ 9:16 am
December 18th, 2005 at 11:59 am
Back to the OP:
Here in the New England is does not pay to be the biggest "baddest" buck. Do you think hunters go after the weak and frail? In Africa it does not pay to be the fastest zebra/ antelope (first to the lioness' ambush).
How many times do accidents (ie chance events) claim the lives of individuals within a population? IOW far from "gloss" Dr. Berlinski's article struck at the heart of the matter. There exist many factors that contribute to the proliferation or demise of an allele, individual and/ or population. Sir Richard would have best said, "the animals that survive long enough and (successfully) mate have the best chance of passing on their genes."
It should also be noted the genes that get passed on do not explain the origin of the organism in the first place. Natural selection can only select from existing qualities. It cannot explain the origins of those qualities. Oscillation of allele frequencies can be explained by NS. Genetic homeostasis can also be explained by NS. IOW NS has all the appearance of a conserving force, just as Blythe told us some 170 years ago.
Comment by Joe G — December 18, 2005 @ 11:59 am