Is there more than one definition of natural selection?

After 300+ posts following Message 1 we failed to agree on the exact meaning or precise difinition of natural selection. I have been saying that NS is the differential reproductive success amongst invividuals of a population. This agrees with E. O. Wilson's defintion (from Sociobiology, 2000, p. 589):

quote:

---

Natural selection: The differential contribution of offspring to the next generation by individuals of different genetic types but belonging to the same population.

---

He does not mention sexual selection, mutation, gene flow, or drift in his definition; he only refers to the differential reproductive success amongst individuals in a population. Is there a cause/effect relationship within the context of NS that is not yet well understood? Do members of this forum think that natural selection is a cause, an effect, or both, as it associates with a microevolutionary event? It certainly does seem causal to me. So why do we still disagree on how to define NS?Does anyone have a better definion of NS than the one I provided?”HM

Here’s a different angle on differential reproductive success”a 'twofer':Nature Online (March 26, 2007) reports Semi-identical twins Discovered:

quote:

---

Researchers have discovered a pair of twins who are identical through their mother's side, but share only half their genes on their father's side.

​

The 'semi-identical' twins are the result of two sperm cells fusing with a single egg ” a previously unreported way for twins to come about, say the team that made the finding. The twins are chimaeras, meaning that their cells are not genetically uniform. Each sperm has contributed genes to each child...This twin turned out to be a 'true hermaphrodite', with both ovarian and testicular tissue.

---

The researchers called the twins chimaeras; so rare, they said, that they didn't to expect to ever see another occurrence.Do semi-identical twins ever occur in other species? I suppose they do, but probably rarely. Co-fertilization, if I may call it that, would seem to be another possible cause of differential reproductive success amongst individuals in a population.”HMEdited by Hoot Mon, : Fixed link

brennakimi, I fixed the link.”HM

WK, I’ll have to concede this one to you. Co-fertilization does not necessarily amount to differential reproductive success; thus it cannot be regarded as NS, or a quirk of it, on that basis. I seem to have strayed a bit myself from what I understand NS to me.I do prefer the gene’s-eye view of evolution and NS. So, from that stand point, I am interested in the frequency distributions of alleles, and how those distributions might change. Given that, I’ll try to back myself out of a corner:Differential reproductive success amongst individuals in a population would mean that not all individuals contribute the same number of offspring to the next generation. Simple enough: if every individual of a population produces the same number of offspring there is no differential reproductive success (i.e., no natural selection). Now, taking into consideration what this means genetically, every individual’s gametes carry their own meiotically shuffled complements of alleles. So, upon successful fertilization, the zygote gets an even half-and-half allele mixture from one sperm and one egg. In this normal arrangement there are two allele-mixing opportunities”meiosis and fertilization. But, in the case of those semi-identical twins, there were three allele mixing occasions: one in meiosis and two in co-fertilization. This suggests to me that, on such a rare occasion, the allele frequencies can be altered (albeit slightly) by this extra mixing opportunity. As such, it would be an additional cause, possibly explaining the change in a population's allele frequencies.In any that case, though, NS remains only a measure of differential reproductive success amongst individuals. Thanks for being sharp on that.”HM

Modulous wrote: I do agree. Modulous, I would venture to challenge your algorithm this way:1. What does “variation in a trait” mean? Alleles are variations of genes, of course, and in this way they ARE the traits. The genes themselves are the not the traits, per se; that job goes to the alleles. For example, I don’t call an eye-color gene a “trait.” I call the alleles of the eye-color gene “traits.” So, the first term in your equation actually pertains to different allele frequencies. The Hardy-Weinberg (dis)equilibrium would seem to account for that.2. What does “heritability of that trait” mean here? Do you mean that certain alleles or their frequencies are more heritable than others? Probably. But all it implies to me is that traits can have variation, even in their heritability, which seems already accounted for by the first factor of your algorithm.3. “Differential fitness conferred by that trait” could be another way of saying “favorability of that trait.” As such, the favorability part seems close to meaning of “selection.” But, of course, selection here seems to be accounted for in "natural selection," on the other side of the equation. Good and subtle point, but confusing. Doesn’t this cast doubt again on the meaning of a "trait"? Isn’t a trait, by way of viewing it generationally, really just a lineage of allele frequencies? I don’t know of any non-heritable “traits.” How would they last long enough to be a trait?Let’s go back to someone’s model of natural selection in an earlier thread, I think it was Quetzal’s. He used a game of marbles cascading through tiers of barriers with holes in them to model NS. The marbles and the holes were of different sizes, so selection, fitness, or favorability was decided on the basis of size. (The author of this model pointed out that its simplicity may not fairly represent the complexities of natural selection and evolution.)So, thinking adventurously, perhaps your algorithm could be revised this way:N = dF x SWhere N is natural selection, dF is the disturbance of the HW equilibrium, S is a measure of favorability or selectivity (selective pressure? information?). I multiply the terms of N for arbitrary reasons, mainly because they don’t seem to be just additive to me, and also in this way the model resembles the structure of Ohm’s Law:V = IR,were voltage V equals the current I times the resistance R. So, the resistance metaphor relates to the tiers, and the current is the movement of marbles through the tiers. What results is differential reproductive succcess ("the voltage of natural selection").Thanks for letting me experiment a little with your model. I'm sure I will learn soon that this was a bad idea.”HM

Hold on a minute. I saw a three-legged dog the other day and it never occurred to me that he is a bearer of "the three-legged dog trait." I really don't think he is. I think all "traits" are genetically determined characteristics. If not, then, by your reasoning, a dead dog suffers from "the dead-dog trait."”HM

Percy writes: Yes. Percy, what leads you to say "...but selection pressures were present just like they always are, and so the process of natural selection was still in force"? There is nothing in your scenario that means or implies that "selection pressures were present." How would you know this if there is no selection? Unless there is differential reproductive success amongst individuals, there simply is NO natural selection. Period.”HM

I've printed out Millstein's paper but I still haven't read it (weather's been too nice today). But I will read it to see what I can learn, and comment later. Good question. NS seems to have a zero or rest state, and it seems to have an active state with variation of activity, but I don't know if a scale of effect would be possibly to draft or measure. Still, philosophically, at least, anything that has a 'differentiality' should be measurable somehow. (I guess many of us have already argued at length on another thread about where we should insert the probe of an NS-detection meter, if there were one.)*”HM*I expect crashfrog will drop by soon to tell me where to insert the probe. Edited by Hoot Mon, : typo

Correct, if you observe it occurring across a population. Percy, you've spun me around with this one. It's your "negative differential reproductive success" that is troubling me. But I'll soldier on. This flies south on me. Are you saying a population can host three states or charges of natural selection: positive, neutral, and negative? I'm not comfortable viewing NS this way. NS can be either active or not; there is no negative activity that has any meaning in the context of differential reproductive success amongst individuals across a population. The trouble with your example is that it assumes NS can be analyzed amongst three individuals. I would be more comfortable looking across a population to do that. If NS can do what you say it does on a population of three individuals, I wonder what role random genetic drift would play in changing that population's allele frequencies. Probably quite large, as compared to the role of NS.”HM

Mod wrote: Yes, I would rather, because "trait," in my lexicon, implies heritability. Otherwise, we're going to struggle with a "three-legged-dead-dog trait," where neither Darwin nor Lamarck would have anything relevant to say about it.”HMEdited by Hoot Mon, : added subtitle

WK wrote: Should I also consider altruistic anti-selection? Not if you care to use the agree-upon definition of natural selectrion. Why do you need to complexify it? Don't anwser; I think I understand. It's that nagging issue of cause v. effect”process v. outcome”isn't it?Maybe so. In this respect, the paper by Millstein you asked me to read”“Are Random Drift and Natural Selection Conceptually Distinct?””is worth the reading. If I get the major thrust of it he is saying that they may be indistinguishable, depending upon whether you view them as processes or outcomes. I think he's right, mostly; cause and effect are not at all the same things. And this casts certain doubt on our efforts to differentiate the roles of NS and drift. (There's more to say about this paper, but I'll postpone it for now.)So we're back to worrying over cause-and-effect relationships. Indeed it may be productive for us to differentiate process from outcome, because many of us here don't bother to make that distinction.”HM

WK, I should have taken more time with Roberta Millstein’s paper”too hasty in getting back to you, I guess, but spring sailing has distracted me. I mentioned just one aspect of her paper that I happen to think is important, and that we haven’t yet discussed enough. So I hurried on from there.I found her paper interesting and relevant to our discussion. She takes issue with J. Beatty’s contention that it is conceptually difficult to distinguish NS from drift:

quote:

---

Again, in Beatty’s account, natural selection is discriminate sampling, or sampling with regard to fitness differences, whereas random drift is indiscriminate sampling, or sampling without regard to fitness . Yet Beatty maintains that these concepts are not conceptually different . However, there are other possible responses to the problem that Beatty raises. For example, Shanahan (1992) argues conceptually random drift and natural selection are the ends of a continuum.

---

She goes on with a rigorous discussion of the concepts, making good points about the differentiation of discriminate from indiscriminate sampling (NS v. drift). She believes, as I do, that they are conceptually distinct. I enjoyed the paper and learned from it. Thanks.Picking up on this concept that views NS and drift as “the ends of a continuum””in others words the “outcome,” and we’re right back to the process v. outcome controversy.I also wonder also about “co-processing” "counter-processing."Assuming they are distinguishable from one another, the five categorical causes of the “outcome” we perceive to be microevolution are: 1) differential reproductive success, 2) random genetic drift, 3) mutation, 4) gene flow, and 5) preferential mating. I view them as processes rather than results, but one process might work on,with, or aganist another. The outcome ultimately must mean that a new allele-frequency equilibrium is established after certain formative processes have taken place. I would call this “microevolution.” But microevolution is not likely to be caused by only one of the known causes. They probably work in combination, which doesn’t make them any easier to differentiate.And maybe there is even “counter-processing.” I take notice in Essential Genetics/A Genomics Perspective (2002, p.523) that Daniel Hartl and Elizabeth Jones conclude:

quote:

---

It is apparent from [their studies] that selection tends to eliminate harmful alleles from a population. How, harmful alleles can never be eliminated totally because recurrent mutation of the normal allele continually creates new harmful alleles. These new mutations tend to replenish harmful alleles eliminated by selection. Eventually the population will attain a state of equilibrium in which the new mutations exactly balance the selective eliminations .

---

How do they define “selection”?

quote:

---

selection In evolution, intrinsic differences in the ability of genotypes to survive and reproduce; in plant and animal breeding, the choosing of organisms with certain phenotypes to be parents of the next generation; in mutation studies, a procedure designed in such a way that only a desired type of cell can survive, as in selection for resistance to an antibiotic.

---

And how about “natural selection”?

quote:

---

natural selection The process of evolutionary adaptation in which the genotypes genetically best suited to survive and reproduce in a particular environment give rise to a disproportionate share of the offspring and so gradually increase in the overall ability of the population to survive and reproduce in that environment.

---

So, by their definition, NS is a process, or maybe an amalgamation of co-processes and counter-processes.Regarding Hartl’s (and others’) experimental work to test the accuracy of predictive models of random genetic drift, it’s worth posting here one graphic representation of this effort: “Figure 14.27 (A) Random genetic drift in 107 experimental populations of
Drosophila melanogaster, each consisting of 8 females and 8 males.
(B) Theoretical expectation of the same situation, calculated from binomial
distribution.” (scanned and quoted from p. 528, Hartl & Jones, 2002)So, at least experimentally in very small populations, drift can be both modeled as a process and calculated as a result. But, in nature, I doubt if the “continuum” would end there.”HMEdited by Hoot Mon, : No reason given.

I do like this argument over sexual selection = natural selection (or permutations thereof). I'm sticking with my side: they are not necessarily an equality. I'll give you this much: nonrandom mating could work as a co-process along with natural selection; it may even work against it as a counter-process (as I discussed in Message 32).I'm, finding less time these days to stay in these discussions. They keep getting deeper all the time. And that's good. But, oh, the commitment...”HM

I will pick up on this notion of co-process v. counter-process, as I discussed in Message 32. I’ll argue that natural selection is a process, along with all the other processes, including sexual selection.I don”t see why sexual selection (or nonrandom mating) needs to be only imbedded in natural selection as one of its relegated features. But I do see how a causal continuum needs to be considered. I’ll take the position that any or all five known microevolutionary causal processes, co-mingling in that continuum, can make a microevolutionary event happen. This may be the most realistic point of view. As Hartl and Jones (2002, see Message 32) have already concluded, mutations may affect natural selection either by enhancing or suppressing it. I could see similar effects”enhancement or suppression of NS”from the other three causal processes attending microevolution, including nonrandom mating. I single out nonrandom mating here because, in and of itself, the process refers only to one way an HW equilibrium can be disturbed.How could sexual selection, as a process, suppress natural selection?If natural selection”which by definition is the ”differential reproductive success amongst individuals of a population’”selected for trait (characteristic) X, it is also possible that, by my reasoning, nonrandom mating could select neutrally or against trait X. Yes, I’ll agree that in either case NS may be affected by the causal process we call sexual selection, but if nonrandom mating somehow suppressed NS enough to neutralizing it, I would have to conclude that sexual selection could be a counter-process as well as a co-process.Sorry, I don’t have any examples. But try this: Imagine a population wherein the good-looking ones mated exclusively with each other (by virtue of traits), but they didn’t reproduce as much (per capita) as the bad-looking ones, who had to settle for other bad-looking ones, but who were coincidentally more fecund. If I put this on a sliding scale, I could envision sexual selection swamping out NS altogether.Why not?”HM

What do you expect from a geezer?btw: Is that all you have to say about drift v. selection?”HM .