As shown repeatedly, “Nothing in Evolution Makes Sense. Period.” Not natural selection, gradualism, human evolution, UCD, tree of life, etc. And just to confirm, let’s look at another one of the nonsensical concepts of “evolution”.
- ‘Divergence of character’ (character displacement or sympatric speciation) postulates: “during the incessant struggle of all species to increase in numbers, the more diversified these descendants become, the better will be their chance of succeeding in the battle of life. Thus the small differences distinguishing varieties of the same species, will steadily tend to increase till they come to equal the greater differences between species of the same genus, or even of distinct genera” (Darwin 1859). Sympatric speciation is hypothesized as “the evolution of a new species from a surviving ancestral species while both continue to inhabit the same geographic region”.
- ‘Regression to the mean’ is the biological law that overrules passive ‘Divergence of Character’. Any homogeneous population can be sorted statistically on various biologic metrics, usually resulting in a Gaussian (normal) distribution that is conserved over time in the absence of major environmental changes (as Mendel first showed; Fig 1&2). ‘Regression to the mean’ is thus the rule that causes the progeny of extreme individuals to be less extreme than their parents. Two outstanding tall parents will have statistically shorter children, and the progeny of the most and least intelligent/strong/aggressive/attractive/etc. will be more average than the parent. Many of the extremes have no descendants at all due to their limitations, and thus their “contribution” to the next generation is simply the average individual.
- In stable environments, population variability is extremely well conserved from generation to generation (Fig 3) as documented by the fossil and many other records. ‘Regression to the mean’ is thus a mathematical necessity without which a passive ‘divergence of character’ would be observed in very few generations (Fig 4). ‘Regression to the mean’ mechanism is incredibly accurate and allows for conservation of traits over thousands upon thousands of generations as observed. Scientists were rightfully surprised that ancient bacteria and many other fossils as well as mummified organisms including cats and monkeys are indistinguishable from their contemporary descendants. At a minimum, the number of organisms that show remarkable stability over long periods (living fossils) invalidate the ‘General Divergence’ theory. Does a limited, ‘Special Divergence’ hypothesis still make sense?
- Observed long term regression is highly unexpected and contrary to ‘divergence of character’ and ‘drift’ hypotheses. ‘Regression to the mean’ operates in the longest term observed, whenever environmental conditions are restored following significant changes that led to adaptive mutations. Most – if not all – organisms are endowed with a limited ‘plasticity’ trait that allows them to retain adaptive characteristics for generations. And yet, when the stimulus that caused the adaptation disappears, these organisms regress rather than maintaining those adaptive traits or accumulating even more diverging ones. Darwin’s finches, the peppered moth, antibiotic resistant bacteria and the domesticated plants & animals – all these and more have been observed to regress to the old mean when the adaptive stressor is removed, thus disproving even the limited, ‘Special Divergence’ hypothesis. These are not coincidences! The regression can happen over a few generations as in most epigenetic changes, many generations, and even the indefinite future if the adaptive stimulus is maintained (such as in domestication). Biologic variability can be compared to a loaded spring – the more it stretches, the harder the pull back (regression to the mean) and the more fragile is the extreme variant population. Domesticated plants and animals show that crossbreeds are resilient, while pure breeds are fragile showing that extinction of the extremes is the default outcome that promotes the ‘regression to the mean’ of the extended population.
- Adaptation neither demands not implies divergence in any way. What about the ‘adaptive radiation’ seen in Darwin’s finches, the cichlids of the African Great Lakes, and others? Is this not ‘divergence of character’? No. The driving force in all these and more is adaptation, not divergence even if “evolution” were true. Organisms just seek survival and, if their built-in yet limited plasticity matches the environmental challenges, these populations survive as variants. Otherwise, they simply go extinct like many others before. The new traits are not ‘divergent’ as shown by all known cases of reversals (as discussed) and none of further divergence when the adaptive stressor is removed. If ‘divergence of character’ were true, adaptive plasticity traits would be cumulative and sticky even after the adaptive stressor was removed, and the more extreme variants would be at least as resilient as the mean. Furthermore, experiments would show increasing variability over time in all research organisms and even more so in the short lived ones like bacteria. There would not be any distinct “species” and organisms would freely undergo metamorphosis (transmutation) into one another. Differential survival and randomness would eliminate all but the “best adapted” allele, therefore the Mendelian conservation of alleles would not be observed. Yet none of these are happening, thus falsifying the ‘divergence of character’ hypothesis.
- Adaptation is “fast and done”, “do or die” by necessity, unlike the supposed “slow and ongoing” ‘divergence of character’. If adaptation is not fast enough, the population simply goes extinct as many others did. The cichlids of Lake Victoria had less than 15,000 years to adapt and are as diverse if not more so than the cichlids in the other, much older African Great Lakes. But they do not need even that much time as the newer aquarium varieties obtained in a few generations show. Most likely, cichlids variants have come and gone throughout the history of all African Great Lakes in short cycles of adaptation. And that is why the cichlid biodiversity difference between a few years (Lake Victoria) and millions of years (other African Great Lakes) is unremarkable. The only remarkable fact is that cichlids have a predominantly Gondwanan distribution showing that in 180+ mil years, they did not adapt to ocean living despite their otherwise high adaptability. This clearly shows the limitations of adaptability and makes it an unlikely substitute to ‘divergence of character’. Darwin’s finches, peppered moths, bacteria, and many other also adapt fast or die as observed. And when the stimulus disappears, they revert just as quickly, and later readapt to whatever new stimulus they face or simply die out trying as confirmed. It is a very good thing ‘divergence of character’ is false, or else antibiotic resistant bacteria and other superbugs would have killed mankind by now as “evolution” falsely predicted.
- Statistical evidence refutes ‘divergence of character’. According to the theory, “when organisms compete for scarce resources, natural selection should favor those individuals that are least like their competitors”. And since organisms always “compete for scarce resources”, the least average members of a homogenous population should always be favored by “natural selection”. If so, the well known normal distribution of any organism dimension (length, height, weight, etc.) should always be under pressure to change. We should see groups of “least like” the average form second, third, and so on normal distributions of their own, thus reshaping the original normal distribution into a composite distribution with several peaks and valleys as in Fig 5. And even that should not be adequate, as any concentration of similar individuals would be disadvantaged according to the ‘Divergence of character’ hypothesis, thus leading to uniform distributions as in Fig 4. However, neither Fig 5 nor uniform distributions are seen in homogeneous populations. Instead, we always see normal distributions. And since we see the normal distribution maintained over arbitrary number of generations and no hint of transitioning to a uniform distribution, the ‘Divergence of character’ hypothesis must be discarded. A trend not supported by several period observations must be discarded as noise artifact. This is the case for all examples considered including Darwin’s finches, the peppered moth, antibiotic resistant bacteria, cichlids, etc. All seem somewhat supportive of the divergence hypothesis over carefully chosen periods, yet the divergence is clearly illusory over longer periods.
- Are the bear of North America not like Fig 5? Yes, but they occupy different geographic regions. They are not homogenous. Indeed, we do encounter subfamilies of organisms, that have normally distributed metrics within the subgroup yet clearly distinct from those of other subgroups. However, where these subgroups overlap, the blend is always geographic and never biologic, meaning we see fewer of one kind and more of the other when moving from one’s territory to the others’ instead of blended characteristics as ‘divergence of character’ would predict. Humans are not different “species” although various subgroups are exclusively vegan/carnivorous, white/black, extra small/large. And domesticated organisms including canids are even more diverse than humans. Are the wild cichlids, finches, mice, and others qualitatively different than humans and canids? No. Then why the different “species”, many of which, ironically, are threatened by hybridization? The unwarranted inflation of “species” that do not even meet the loosest definition of reproductive isolation has the sole purpose of perpetuating the myth of ‘divergence of character’.
- Multimodal distributions in homogenous populations are not due to ‘divergence of character’. Indeed, bimodal distributions (Fig 2) and multimodal distributions are not uncommon in homogenous populations. However, these are due to the discreteness of physics in general and biology in particular, not due to ‘divergence of character’. Male and female populations are not diverging from one another and various alleles are in long term cyclical equilibrium as shown (spring model). ‘Drift’ is often invoked as a mechanism of ‘divergence of character’. This is wrong because ‘drift’ explains nothing as it is either aimless noise or due to adaptation and environmental change. Yet, as shown, adaptation is in no way ‘divergence of character’. In addition, the stable coexistence of several distinct variants within a homogenous population shows “gradualism”, “survival of the fittest”, and “natural selection” to be false because the alleles responsible are themselves distinct (no “gradualism”), they all “survive”, and neither is “selected” for or against.
- Darwin worried about regression to the mean for the wrong reasons. Namely, if blending inheritance (Darwin laid an egg) was true, then natural selection could not be true. Darwin puzzled over this a lot, but ended up with nothing satisfactory. Then Mendel showed that inheritance is discrete, not blended. Mendelian Inheritance Tables (see Punnett squares / Hardy-Weinberg equilibrium) show “probabilistic traits conservation” and thus disproving ‘divergence of character’ (at least as byproduct of reproduction) as well as dismissing “gradualism” (another one of Darwin’s unsupported claims).
- When entire populations split, do subgroups diverge from one another? This is not how ‘divergence of character’ is supposed to work.Descendants are supposed to diversify within the homogenous population. Furthermore, populations split by environmental conditions simply adapt to the new environment and for as long as those conditions allow. Adaptation is the driving force with no ‘divergence of character’ anywhere in sight. Island biology is the most diverse because islands are isolated and have many microenvironments. However, island variants are close descendants of their original colonists, showing that no divergence ever happened. Their risk of hybridization is high, disproving the “speciation” claim. They are also fragile examples of the extreme stretched biological spring model discussed, and will likely go extinct if at all stressed and when interacting with mainland.
Summary:
1. ‘Regression to the mean’ is the biological law that overrules passive ‘Divergence of Character’
2. In stable environments, population variability is extremely well conserved from generation to generation
3. Observed long term regression is highly unexpected and contrary to ‘divergence of character’ and ‘drift’ hypotheses
4. Adaptation neither demands not implies ‘divergence of character’ in any way
5. Adaptation is “fast and done”, “do or die” by necessity, unlike the supposed “slow and ongoing” ‘divergence of character’
6. Adaptation has limited powers and is thus not a substitute for ‘divergence of character’
7. ‘Divergence of character’ hypothesis would lead to uniform rather than normal (Gaussian) distributions as observed in homogenous populations
8. A trend not supported by several period observations must be discarded as noise artifact
9. The unwarranted inflation of “species” that do not even meet the loosest definition of reproductive isolation has the sole purpose of perpetuating the myth of ‘divergence of character’
10. Multimodal distributions in homogenous populations are not due to ‘divergence of character’
11. Mendelian tables show “probabilistic traits conservation”, disproving ‘divergence of character’ (at least as byproduct of reproduction), as well as dismissing ‘gradualism’
12. Island biology proves adaptation and the biologic spring model while disproving ‘divergence of character’
13. What’s in, what’s out? IN: ‘regression to the mean’, ‘adaptation’, coexisting variants, long term stability, spring model, normal distributions. OUT: ‘divergence of character’, gradualism, drift, speciation, uniform distributions, “natural selection”, “survival of the fittest”, “evolution”.
Links:
https://ucmp.berkeley.edu/bacteria/bacteriafr.html
https://www.sciencedaily.com/releases/2019/10/191018112136.htm
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3285564/
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3352989/
http://www.galton.org/essays/1880-1889/galton-1886-jaigi-regression-stature.pdf
https://en.wikipedia.org/wiki/Character_displacement
https://en.wikipedia.org/wiki/Sympatric_speciation
https://www.bionity.com/en/encyclopedia/Character_displacement.html
https://biologydictionary.net/divergent-evolution/
https://en.wikipedia.org/wiki/Cichlid
https://biology.stackexchange.com/questions/41982/regression-to-the-mean-and-evolution
It’s clear despite your willful ignorance: “the small differences distinguishing varieties of the same species”… and… “sympatric speciation” these days.
If fake ignorance is your defense, then “divergence of character” means nothing to you and you’re just wasting everyone’s time commenting here.
So no example? Maybe Jock knows what he meant.
It’s your burden to give such examples. There aren’t any. And what does “human time-scales” mean? I’d say at least 5000 years (so plenty of time) given we have artifacts and mummies from that time that show no “evolution” whatsoever. I explained:
“Scientists were rightfully surprised that ancient bacteria and many other fossils as well as mummified organisms including cats and monkeys are indistinguishable from their contemporary descendants.”
Your point is wrong – we do not base our science on wishful thinking and that’s why “evolution” is not science. The true point – the one you keep missing is: “A trend not supported by several period observations must be discarded as noise artifact.”
Exactly what “basic genetics” support your views? And how?
Not what you asked for. Go check.
What’s your formula?
Your point being? We know regression to the mean to be fact. There’s no need for your model to prove it.
See? This is not part of the sexual reproduction.
You only have this due to the hybrid reproduction. With pure sexual reproduction you never have this except only by design in the lab environments.
Why should M mate with DD instead of third party? Why D1 with D2 instead of third party?
It’s mitosis. Which is a necessary part of ANY life-cycle, sexual or otherwise.
What utter rubbish! “Hybrid” refers to mating between distinct species. Not the case here. I am describing what happens in the wild. In ‘lab environments’, mating type switching is usually disabled, because it’s a pain in the neck.
Because they are touching each other already, and in the wild there will often be nobody else nearby.
Again, I have been describing what usually happens in the wild. Go educate yourself, if you dare.
There’s a metric called the “vomit 50 index” of how toxic possible prey are to birds. Some truly dedicated researchers forced birds to eat various potential prey, and counted how many times the bird vomited it up.
I have proposed a similar “biologist 50 index”, to evaluate such matters as exactly when life emerged from proto-life, or exactly when our evolving ancestors reached the point we’re regard as modern humans. To apply this metric, we put 1000 biologists into a time machine, and have them move through time taking samples. We can say that life began when 50% of them decided an object was alive, or when 50% of them say some hominid species became modern humans.
I would expect that as time passed, the percent of biologists deciding something was alive or human would gradually increase. Hopefully, 100% would decide modern humans ARE modern humans, and actually alive, if they took their samples today.
I don’t know whom you are quoting here, but I’d ask for an operational definition of “ancient.” I have seen estimates of the emergence of modern humans varying from as few as 200,000 years ago to as long as 2 million years ago. THAT is what I’d consider the human time scale. 5000 years is an evolutionary eyeblink for cats or monkeys.
I have read (but certainly not exhaustively) that most evolution involves speciation, that most speciation occurs rapidly (in geological terms) and locally. This makes the observation of evolution exceedingly rare in the fossil record, because you need fossils from exactly the right time and place, AND multiple samples from that bullseye. The typical longitudinal pattern, entirely consistent with fossilization, is for some species to show up, hang around without change, and then vanish.
Enough digging has been done (I think mostly for oil) to follow the changes in some species that had shells and fossilized easily. And some of these species show no changes for periods upwards of 15 million years – after which they simply vanished from the fossil record.
ETA: Also, generally not much is preserved for very long of any species. Imagine fossils of jays – could the paleontologist distinguish from the bones that they had fossils of blue jays, canyon jays, scrub jays, etc.? Could they even distinguish a jay from many birds of similar size?
And you have evidence that those introductions took the form of a single pair of mice?
In your narrative, when did this couple live? Did they live at the same time?
You seem reluctant to address the issue of when your first couple lived. Why? Are you scared?
How old is the earth Nonlin?
No, you’re lost! 😛 They weren’t a ‘couple’, they were two individuals living a long time apart who happened to contain the common ancestor sequences of fragments of our genome – those fragments which are, respectively, passed on only through the female and the male line.
Mitochondrial Eve was the ancestor from whom all mitochondria derive. ‘Sperm’ (actually, Y-chromosome) Adam contained the most recent ancestor of Y chromosomes. They never met, let alone got it on. The rest of the modern genome traces its ancestry to other individuals at other, generally prior, times.
No, you started it! 😛 You insist that sex is ‘for’ combining varied inputs to generate varied outputs, but that can’t work efficiently given the genetic (not “evolutionary”) problem of inbreeding. You insist you can argue this topic while remaining clueless about genetics, so my task is made harder by your obtuseness.
Inbreeding is a very well-known and well-characterised problem for populations starting from a narrow gene pool. It is hardly up to me to ‘prove’ it.
If Adam and Eve have children, the only partners they can have are their brothers and sisters, or mum and dad. That’s pretty inbred. At the next generation, cousins can marry cousins, but normally such cousins only have one or two grandparents in common. These children-of-inbreds have all grandparents in common. And so on. The same few genes are shuffled round, more and more of the original variation is lost, and the population looks more and more like a clone. Variation is further squeezed and the distributions of character states are thereby narrowed. No wonder they’re so depressed.
Of course, mutation will eventually restore some variation, if you wait long enough. But that, my blinkered and obtuse friend, is evolution.
A commenter at talk.origins expressed the ‘preservation’ problem succinctly: “where are all the fossilising passenger pigeons?”.
What is striking, too, in a world where evolution is for-argument’s-sake ‘not true’, is how we can drill down through some relatively narrow strata and even see the ‘morphological stasis’ critics point to as evidence against. In a world where morphological evolution is not just sporadic but non-existent, I’d wonder what accounts for this banding?
That’s the beauty: It is your model. You wrote it. I only changed a tiny detail to better reflect the contribution of the heritable and non-heritable component of variation. That can be simply expressed as:
where P is the phenotype of an individual (individual cells in your spreadsheet). G is the genotypic value associated with its genotype. If you perfrom the tiny correction I suggested, the first colum represents G. It models the transmission of heritable variation along your family lineages (in the rows). Finally, E is the environmental deviation, which is non-heritable and therefore has to be recalculated every generation. We assume genetic and environmental variation are independent.
It is better not to think of models as “proof”, but rather as formalizations of verbal models. Just like a nonsensical verbal argument, a nonsensical mathematical model proves nothing. The adequacy of the model is judged by how well it describes real world phenomena.
Happily, the corrected model in all its simplicity recapitulates several salient features:
* The character distribution is stable across generations (which you wanted to show)
* Variation in the phenotype is (partly) heritable
* There is regression to the mean
Now, IF you accept this model as correctly describing inheritance of quantitative traits, THEN we can use it to explore whether regression to mediocrity will push back the population to its original mean after a selection episode. Heck, you can even explore that yourself. You have the model on your own computer!
Maybe this will help you, nonlin.
Look a Corneel’s scatter diagram. Better yet, create it yourself, by editing your model as he described.
Note that the data points that are “extreme” in generation 3 have a strong tendency to be less extreme in generation 4.
This is “regression to the mean”.
Now, flip the arrow of time.
Look at the most extreme points in generation 4. Note that they have a strong tendency to be less extreme in generation 3.
This, too, is “regression to the mean”.
“Regression to the mean” is a statistical truism, for any comparison of repeated measures — in fact any situation where two measures are partially correlated.
It is not some kind of magical force that reduces variation generation-over-generation. It does no such thing.
I do not want to engage with nonlin.org’s bald assertions, but just let me mention what quantitative genetics theory says about regression to the mean. Each generation, if there is natural (or articificial) selection, the mean of the population can shift as the gene frequencies change. Regression occurs, but it is to the new mean in each generation. So regression to the mean is, as Corneel mentioned far upthread, not a barrier to change.
Joe Felsenstein,
I’m seeing no reason to change my impression of a “bumblebees can’t fly” scenario. “My model proves evolution wrong because it regresses to the mean” but observe reality… Variation exists and persists, selection occurs. Bumblebees do fly.
Alan Fox,
Sure, but us theoreticians then are faced with the question of whether there is something wrong with our theory and with the types of models we make. And in the case of regression to the mean the answer is no, its explanation is easy-peasy, just the ordinary behavior of multivariate normal statistical models that have environmental effects and dominance deviations in them.
Similarly for bumblebees.
No? Why?
Here is what we have instead… 😉
Budding is asexual reproduction by any standard. Not what we were discussing!
This started with “descent with modification” which is indeed as retard today as it was in 1859 for ALL sexual reproduction. Yes, people knew back then that the product of sexual reproduction is a ‘mix’ AND NOT a ‘modification’.
BTW, there’s no monopoly on the word “hybrid”.
Pay attention. We were talking about human time-frame verifiable observations. Also, this whole thing about “emergence” and “evolutionary”, “speciation” is bogus.
Don’t you forget about this, Corneel. The world is still waiting for your wisdom.
Need not be. A bottleneck effect would do.
Not my job or inclination to speculate… like this guy:
How the heck would you know? Perhaps they mated with some lions instead of each other. Did you hold their candle?
Citation? You’re making up stuff.
For ever? Can you prove this BS?
Now you’re talking. Of course. Only that is not “evolution” because you still have humans, or mice or whatever. See bottleneck comment above.
So you hold genotype constant for ever? And you inherit directly from your ancestors as in $A vs A?!? Do you understand this is not how it works?
Then you have a constant a random variable and you’re delighted that the distribution is preserved?!? Wow! Peekaboo to you my little friend.
And what about P(-1)? Is your height not related to that of your parents?
What “selection episode”? Who does the “selection”? Will you tell this to the trilobites that were virtually unchanged for 300 million years?
After my 10 gen simulation do you understand what a MIRACLE this regression to the mean is that we end up no different than Otzi and the Egyptian mummies? And how strong of an evidence against “divergence of character” and “evolution” this is? Of course not.
FALSE. The arrow of time shows gen 3 NOT AT ALL a function of gen 4.
So what are you afraid of?
Of course the mean can change – it’s not rocket science. Yet how do you account for its eventual reversion as in Darwin’s finches, the peppered moth, antibiotic resistant bacteria, domesticated plants & animals?!? All these and more have been observed to regress to the old mean eventually. Never to “diverge” to something else.
No. This is just one of MANY nails in “evolution’s” coffin. And you’ve been exposed to a few of them.
Maybe at some point you will come down from you ivory tower, smell the roses, and see that your theories are totally unrelated to reality. But then again… “It is difficult to get a man to understand something, when his salary depends on his not understanding it.”
This is not the way to convince anyone you have a genuine point to consider.
Speaks volumes! 😉
nonlin.org has already, in effect, conceded that his point about regression to the mean being a barrier to evolutionary change was wrong. But nonlin.org says that “This is just one of MANY nails in “evolution’s” coffin. And you’ve been exposed to a few of them.” Which amounts to an admission that the point nonlin.org made about regression was wrong.
Of course individuals inherit from their parents, not from their ancestors. But since there is no mutation or recombination in the model, the genome is inherited faithfully indefinitely. That’s why we can use this shortcut. I see now how that confused you. If you prefer, go ahead and program it so it is inherited from mother to daughter each generation: It won’t affect the outcome.
From the OP: “In stable environments, population variability is extremely well conserved from generation to generation”. Isn’t this exactly what you wanted to see? I have just presented a more parsimonious explanation than you did. No need for mysterious unknown forces tugging and pushing the distribution.
LOL! Care to guess how many species of trilobites were named, based on differences in morphology alone?
Not any more. Joe is enjoying his retirement if I am not mistaken.
They lived at different times. If you’re insisting that evolutionists are talking about a ‘first couple’, you are talking shite.
It’s what you’ve said. You’re arguing your ass off with Jock because there would be ‘no point’ to sex if there was no variation.
Not for ever, no. As I say, eventually mutation will add novel variation. But that’s evolution.
It is evolution. “They are still mice” indeed. Hahahaaaa Creationists are such fun.
What’s all this about bottlenecks? What do you need a bottleneck for? You were on about ‘mouse Adam and Eve’ before. Now you say ‘bottleneck’ as if it makes much difference. Why do you need to shrink the starting population at all? Is it because they took a 40-day sea voyage?
I was originally trying to find out where you thought variation originated – whether it was Created or somehow added later. Not having much luck getting a straight answer.
<spittake>
Say what?
Of course they are! Both in Corneel’s and your versions of your model. And in any correlation != 0.
Nonlin is only talking about quantitative traits. And as he asserted upthread, evolution has ‘no connection to genetics’ (guffaw!). So talk of allele frequencies is as the buzzing of insects on a summer’s day.
Nonetheless, I’d like to know what the ‘mean’ of a polymorphic locus with (say) 4% A, 37% B and 59% C is, and how it is regressed to following a perturbation to a new mean – especially if, as could easily happen, one of them is lost to the population, or a variant D appears.
That alone means that there’s no point to even trying to talk Nonlin out of his/her mistakes. Nonlin cannot understand the most basic stuff. We have to wait until (s)he decides to actually have an education and gone through it. I doubt that’s going to happen though. Remember, Nonlin thinks, for example, that assumptions cannot be tested, and no amount of evidence to the contrary will be acceptable to her/him once (s)he takes such a ridiculous position. Nonlin declares it so, and then, by the power of her/his words, it is so.
You asked for the proof that regression to the mean would not stop evolution. Since you have built a model, you can simulate this yourself now. Either use your own model, or the modified one I suggested. The latter will lack the jolly “plumpudding” effect.
Step 1) Choose one of the generations as parents. Calculate the phenotype mean (
).
Step 2) Apply a selection step. I suggest you do truncating selection of the lower 50% (I always took examples with selection for lower height to tease my students, who towered over me). Rank your individuals and remove the upper half. This simulates the introduction of a novel predator that only eats large nonlins. Therefore, it is imperative that you make “aargh” noises as you delete them.
Step 3) Calculate the mean of the selected parents (
). Now you can calculate the selection differential:
Step 4) Breed a new offspring generation from your selected parents. Calculate the offspring mean (
). Make a parent-offspring plot to verify there was regression to the mean. Iterate this process as many times as you deem necessary. I predict the new mean will be stable after the first generation. You can now calculate the response to selection:
If your hypothesis is true, your population will have remembered its original mean, and regression to the mean have will driven it back to its original distribution. The response to selection will be zero. OTOH, if standard quantitative genetic theory is true, the heritable part of the selection differential will have been retained. This is expressed by the breeder’s equation:
where
is the heritability of the trait. If you draw your genetic and environmental variation from the same distribution, the heritability is 0.5 as half of the variation is genetic.
Have fun, and let us know how it went!
This time you’re right. Apologies.
Huh? Citation?
You mean this? If so, your conclusion is illogical.
Nope. You reacted by drawing attention to others of your points, rather than defend your point about regression to the mean being a barrier to evolutionary change. (Your alleged examples of reversal of change are not relevant to a barrier to change at a per-generation level).
Now be good and specifically address Corneel’s mathematical model.
But is the genome constant? What about mutations?
More importantly: what are you trying to do? Coming up with n models that fit a set of data is not hard.
We don’t need a model to duplicate what we see experimentally unless said model has some other purpose. Which is…?
I already explained in the OP: “The unwarranted inflation of “species” that do not even meet the loosest definition of reproductive isolation has the sole purpose of perpetuating the myth of ‘divergence of character’.”
Why did you ignore?
Did I? The topic is “divergence of character myth” (narrow), not “evolution” (broad). And why would I need a random model when we can observe “no divergence” with the naked eye. Are you even on board with what ‘divergence of character’ means to Darwin and if it is same as the very theoretical “sympatric speciation”?
My model had a specific purpose and for that it works great. To demonstrate: “‘Regression to the mean’ is the biological law that overrules passive ‘Divergence of Character’” I would not use it for something else (say space travel).
This is stand alone silly. What has this to do with anything REAL happening in nature?
So you’re talking about human breeding (intelligent design). OK, not what happens in nature, but an interesting case. And we don’t need your random model as we have experimental results: over-fishing and over-hunting of the prized individuals (but not extinction!) and their subsequent [indeed] slow recovery when restrictions are imposed.
No. As per above, I only refocused the discussion on “divergence of character” (this narrow topic) and away from “evolution” (broad).
And this means what?
It’s well known that correlation is not causation. It’s also well known that the height of your parents is not dependent on your height. Why do I have to re-state the obvious?
Going back to an older tangential topic:
No, I don’t seek a degree in S. Cerevisiae (is that a crime?), but it is intriguing that they would use sexual reproduction only when stressed and yet the result would be basically identical to that of budding anyway. Why and why?
So you don’t know but enjoy guessing. And logically speaking, who would they mate with?
So no citation? You have a habit of misunderstanding so go find the citation.
No. That is not “evolution”. We talked about that. Remember?
If you don’t read, you ask silly questions. Go read.
Because it’s not my focus and I don’t have a satisfactory quick answer. Instead, proving “divergence of character is a myth” is the focus of this OP.
Paragraph 9. discusses your question. Go read.
Let’s do a poll:
Who else thinks the assumptions of a model can be tested (proven) WITHIN said model? And if so, how would you argue this?
Please, try to unpack your airtight logic there (that seems to imply that mitochondrial eve would be the only human on earth, Lulz!), so we can have a laugh
Dunno. Because we enjoy watching you make basic math errors repeatedly?
It’s also well known that the height of your parents is not independent of your height.
The scatter diagram below is an example of the outputs of YOUR model, demonstrating that the values in gen 3 are strongly dependent* on the values in gen 4. If you don’t believe me, create the plot yourself in your model…
Ack!
Diploids sporulate when stressed (typically nitrogen starvation), producing haploid spores, which can lay quiescent until conditions improve.
Haploid yeast switch mating types and mate (have sex) to get back to being a diploid. So no, the “result” is not “basically identical to that of budding anyway”. These are two distinct parts of their life cycle.
Read Allan Miller’s paper to understand why.
ETA1: I admit to being a little naughty with ‘dependent’ here, but I’m trying to get him to see the relationship between y-on-x and x-on-y…
ETA2: To clarify re the yeast: the diploid that results from grandpa-granddaughter and brother-sister mating DIFFERS from the original diploid that sporulated: specifically, the original diploid might have a fair amount of heterozygosity…
* as in “correlated”…
You imply there was nobody else.
When this this couple live, the sole two humans on earth?
Nonlin.org,
You are all over the place again. To clarify my position: I took issue with the claims you made in paragraph 2,3 and 4 of your OP. Specifically with the following claim:
You have made clear previously that by “regression to the mean” you mean statistical regression to the mean, as observed in parent-offspring regression plots:
Summarizing, I understand your claim to be that because of “regression to the mean” populations return to the original mean upon removal of the selective pressure that caused them to adapt (the “loaded spring” model)
In my modification of the model you presented, regression to the mean occurs in all generations. However, in my hands, when I apply a transient selection step, the new population mean is stable. Again, I urge you to check this yourself. This strongly argues against your claim that character reversal is a consequence of statistical “regression to the mean”.
You have invested in building a model, which I truly appreciate. As said before, I like it when people get their hands dirty, and I have rewarded that by investing some time in your model myself. Now I expect you to not pull your hands from your own model when it starts showing things you don’t like. You thought it was good enough to support your position in the OP, so it should be good enough to test your claims as well.
To answer your remaining points:
Since you only simulated 10 generations, the effect of mutation is negligible.
Neither is coming up with nonsensical verbal models, but that doesn’t appear to stop *certain* people around here. Models have the advantage that they force us to make some of our assumptions explicit, and they show us how those assumptions shape the outcome.
…gaining insight into the processes that we are modeling. Have you figured out what causes “regression to mediocrity” yet?
Size-selective predation is real.
Unless you are in the habit of constructing cattle in your garage, I have a hard time seeing the connection between breeding and design. Anyway, the breeder’s equation is a generic formula that can be applied both to instances of artificial and natural selection.
I agree. This is a frequent assertion of Design advocates, that animal or plant breeding is design. But when a farmer selects for (say) high milk butterfat content they don’t know in advance which genes will have changes of gene frequency. They have only the target phenotype and not a detailed “design” showing how this is to be achieved.
The distinction seems to escape Design advocates.
Cut out the BS. Of course it is independent. The relationship is unidirectional. Just because yours is dependent, on them doesn’t mean theirs is dependent on on yours. Same childish error as with “mutation = evolution”. Try writing this in code and see it’s not the same as “evolution = mutation”.
“correlated ≠ dependent”. Are you nuts? “Evolution” will not survive with stupid equivocations like these. Won’t survive anyway, but that’s another matter.
1. Your explanation sucks. But never mind.
2. I explicitly said “no degree”, so won’t read your link.
3. Haploids “mating” is definitely not sexual reproduction as we know it.
Yes, that’s what mitochondrial Eve means – the first (only) woman that gave birth to mankind. And she needed one and only one Adam.
We can add statistical dependence to the long list of things Nonlin doesn’t understand
OK. There are two aspects of regression – the gen-to-gen as in the model and the long term regression to some long term mean as observed.
I was wondering about your “selective pressure” deal. You should recognize this is not “divergence of character” as per Darwin and “sympatric speciation”.
Obviously, if you chop off a section of the distribution, the mean changes and then only slowly drifts back to some long term mean provided constant environment.
You keep asking me to do some work. Why don’t you build your own model – but make sure it’s realistic – and then present your conclusions?
As discussed, you can model whatever, but is it realistic? Obviously yours isn’t.
I didn’t say “statistical regression to the mean”. What does that even mean? We can model regression statistically from gen-to-gen, but that doesn’t mean we understand in detail the phenomenon.
I already addressed this several times above and previously. Do I need to repeat again and again?
Let’s talk about that. What exactly is “nonsensical” about it (them)?
I fully agree. “We’re not so different you and I”.
See, that’s where we disagree. Models by themselves don’t do that as they’re just “curve fitting”. Now if said models could predict something beyond their scope (and not just an extrapolation) and that proves to be true experimentally, then maybe they become more trustworthy. But “gaining insight into the processes” I’d say is vague and rather dreamy.
Therefore, no, I have not figured out what causes “regression to the mean” and neither did you. But remember, my purpose is not to nail regression, but to disprove “divergence of character”. And that I did!
No. You just interpret what happens through your fantasy. But predators just wants to eat with minimum effort.
“Designer dogs” means anything to you? Design doesn’t have to mean “from scratch” – a common mistake of those that don’t understand how design works. In fact, design is [almost] never from scratch.
We talked about so-called “natural selection” and there’s no such thing as demonstrated. One of them topics you went silent on, only to start again as if you haven’t learned a thing. Which is probably what you intentionally did.
What non-designers don’t understand is that design does not require a deep understanding of the underlying physics. Farmers have designed farm animals for generations without knowing anything about genes. And had they known, they would not have been any better at designing. As proof, GMOs are not even mainstream yet: “The first genetically modified animal to be commercialized was the GloFish (2003) and the first genetically modified animal to be approved for food use was the AquAdvantage salmon in 2015.”
“Natural selection” is just a shorthand way of saying that the environment, by rewarding some traits over others, is doing the designing. What we call artificial selection is humans manipulating the environment by rewarding some traits over others. From the viewpoint of the dog or goldfish, it doesn’t matter. All that matters is that some traits survive better for reasons they can’t know. What nonlin is trying to argue is that selection WORKS when people do the selecting, but somehow doesn’t even EXIST if the selection is done by something else. A blind spot of impressive scope!
What characterizes evolution is that design is never from scratch. Selection can only work on what’s available to be selected. And because of this, it’s possible to trace ancestry and common-ancestor relationships. If design were from scratch, these relationships need not exist. Indeed, reproduction as we know it would be impossible if each organism emerged unique from the brow of the great Designer In The Sky.
Other members of their respective populations. They are only common ancestors as far as mitochondria/y chromosome, respectively, are concerned, and rates of change put them tens of thousands of years apart. The rest of our genomes come by descent from other individuals.
If it’s not what you think, that’s fine.
It is evolution under the commonly accepted definitions of the term. Darwin used ‘descent with modification’. Mutation is a modification. Dictionary.com says “change in the gene pool of a population from generation to generation by such processes as mutation, natural selection, and genetic drift”. If you have your own private definition, no-one is under any obligation to use it.
I’ve read, and I still don’t know why you need a bottleneck. People with a point to make generally make it. Squirmers do what you just did.
It is entirely relevant, nonetheless. Pulling two threads together, you have stated that sufficient variation can be generated by mutation from a starting couple to provide the mutations subsequently responsible for the adaptive radiations of cichlids and Darwin’s finches. That’s awfully close to the evolutionary position.
It doesn’t address my question. When you look at a population at a given moment, how do you know if your ‘multimodal’ distribution is the mean to which they have regressed, or the perturbed mean? Mechanistically, how is the mean restored, particularly when variants are lost or gained?
And, approximately, how long ago was that?
Hey, where did y chromosome Adam get his mitochondria from? 🤔 🤣
The equals sign there simply served as ‘is’, without any intent to suggest that mutation is the whole of it. It wasn’t algebra; mutation and evolution aren’t variables. “Train = transport” would not mean that the only means of transport is the train.
So now we suddenly have TWO types of regression to the mean? LOL! Could you at least concede that the “gen-to-gen” type cannot reverse or stop evolutionary change? Thanks!
Obviously, the mean stays exactly where it is if the change is heritable. Why on earth would it drift back?
Because, given that you reject the results that came from the model YOU YOURSELF built, there is not a chance in hell that you will accept mine. I put in quite some work here already. Now you need to first learn the theory I outlined so you can attack the actual arguments instead of your army of strawmen.
Why not? “My” model correctly models the contribution of heritable and non-heritable components of phenotypic variation, has a stable distribution and shows regression to the mean. Tell me exactly which part is not realistic and why.
Nonlin, WHY did you build that spreadsheet? Was it not so you could demonstrate what would happen if there were no “regression to the mean”?