Adaptations correspond to physiological ranges (reaction norms, somatic adaptations). According to the theory of facilitated variation, such dynamic physiological restorations of the phenotype in response of variable environmental conditions are the outcome of genetic constraints (e.g. plasticity and robustness of developmental pathways). Therefore, when somatic adaptation occurs, exposing the phenotype to different selective conditions, physiological ranges can be "easily" shifted (i.e. their evolutionary shift is facilitated) by mutation, or genetic reassortments from the existing variability in the population (Baldwinian evolution). In other words, one of the key characteristics of adaptations would be their evolvability, or to say this with the words of Gould and Vrba (1982), "cooptability for fitness". Evolvability can thus be strongly conserved at the level of core molecular processes. Adaptations would then be selected to be both physiologically adaptable, i.e. to function in "a range of ways" in response to changed conditions ("dynamic restoration" or somatic adaptation), and to be evolvable. In other words, the "cooptability for fitness" would be under selection.
In my view, this idea implies that all adaptations at the organismal level should be partly selected to be "preaptations" (sensu Gould and Vrba 1982, Paleobiology), i.e. structures that retain the potential to enhance fitness (adaptive function) in variable conditions. Gould (2002, the structure of evolutionary theory) suggested that this selection should act at higher hierarchical level (species selection). However, the fact that adaptations that are selected in specific conditions are also selected to be physiologically modifiable, or to function in “a range of ways”, would make them likely to have fitness-increasing effects (aptations) that are not those they were selected for during their historical genesis, i.e. to become exaptations. Could this be a bet-hedging strategy also selected at organismal level?
That is, exaptations would also have a non-random origin (contra Gould & Vrba 1982), while this does not rule out the possibility of non-aptations as a possible source of exaptations. In this scenario, exaptations from nonaptations (spandrels s.s.) would be less frequent than exaptations from previous adaptations. Note also that the measurable adaptations are a subset of the extant ones, due to the overall scarcity of available historical data.
I will often modify my comments.
dear, interesting view, but untestable? How do you measure the extent in which any given "adaptation" become a "preaptation"? How can you be sure that any given trait in a paleo-species is an adaptation? Just asking....if you can device how to do such a thing it will be very interesting!
You are spot on.
Evolutionary reconstructions are always reconstructed histories of events.
We may not be able to test the theory looking at the outcomes of the process (structures), but we can test the expectations of the theory on the structures with a correlative (e.g. chi-square) approach, providing observational support.
Of course we need sound hypotheses on the status of these structures: exaptations of the preaptation type, exaptations of the nonaptation type, adaptations (apomorphy), that derive from robust phylogenetic reconstructions and (more infrequently) direct historical inference.
My opinion is that making broad generalizations about evolution, selection, adaptation etc. is rather pointless. Evolution in mammals over several million year timescales with effective population sizes in the millions, is quite different than evolution of RNA viruses over hundred year timescales with effective population sizes in the billions.
With any organism, whether a very tiny RNA virus genome or a very large mammalian genome, there are many different types of selection pressures and "fitness" operating simultaneously. The types of genetic variety that will help butterflies survive a drought or change in weather are different than the type of genetic variety that will help the same butterflies survive increased predation by an expanding population of birds, or a new bacterial or viral disease that attacks them.
Within a single species of butterfly there will be different selection pressures on the caterpillar larval and adult stages.
Likewise for a virus such as HIV-1 there are selection pressures within a single infected host individual that are very different than the selection pressures at the tie of transmission to a new individual and before the immune system of the newly infected host kicks in.
When speaking of random or nonrandom events, we first need a clear definition of what "random" means. For example with HIV-1 we know that G to A mutations happen more than 20-fold more often than G to T mutations, but the HIV genome does not become depleted in G and rich in A over time.
Anyway, the point is that we know of hundreds of simple systems such as bacteria and viruses or phages in lab cultures where we can very accurately observe evolution in human timescales. And we also have observations of vertebrates and other organisms that left fossil records over hundreds of millions of years. So it is possible to discuss very specific details, rather than making broad generalizations.
Brian, observing or testing general patterns should be as feasible as observing or testing specific patterns, either inductively (from the latter to the former) or deductively (from the former to the latter).
Of course, testing the generality of any pattern must be based on specific cases.
Reasoning in the Gould's perspective, since gene mutations are not mathematically random: what could be expected as random is the effect of genetic variability when it is "injected" into the higher organismal hierarchical level. That is, genetic variability is "random" relative to the realised evolutionary pathways at the organismal level, because it is "undirected towards organismal adaptive states". Of course, the differential plurification of e.g. different variants of transposable elements within the genome is non-random, since it is also the result of selection, at the gene hierarchical level. Groups of phyletically related organisms can evolve (populations), as a result of the work done by selection when fueled by the variability produced at the gene level.
Similarly, variability at the organismal level (e.g. flexibility or phenotypic plasticity, reaction norms, physiological tolerance, geometric morphological complexity) that occurs as a byproduct of selection (at the organismal level), is a "random" source of variability relative to the realised evolutionary pathways at the higher hierarchical level of the species, because it is undirected towards species adaptive states (differential survival of species during speciation). That is, also in this case what it is expected to be random is the effect of organismal variability when it is "injected" into the higher species hierarchical level. Groups of phyletically related species (clades) can evolve as a result of the work done by selection when fuelled by the variability produced at organismal level.
I agree with Gianluca in his point of view about non-random origin of exaptations as the subset of adaptations and of adaptations in the large. I disbelieve in the randomness of any of processes in evolution. If i don't see a phenomenon, it doesn't mean that pheomenon doesn't exist.
However, I was proposing a way to test hypothesis that they are more frequent. We just need robust time-calibrated phylogenies and reconstructed ancestral characters.
At the organismal level, the huge "unused" variability that is present at the genomic level (junk DNA) is non(ad)aptive. However, the evolvability of organisms strongly depends on such nonaptive variability, that can be exapted in different selective regimes. At the species level, the variability at the organismal level is similarly nonaptive. But the evolvability of species strongly depends on this variability, that can provide raw material for selection among species in different selective regimes.
My point is: is it possible that the accumulation and maintenance of such presumably nonaptive "junk" is actually under selection at the higher level?
I think of the bet-hedging strategy of angiosperm seed banks...
In an unpredictably changing selective regime, what likely happens to the genotype (genetic variant at the organismal level) that gets rid of too much genomic junk? What happens to the species that gets rid of too much variability among individuals?
In fact, this is what Gould (2002, p. 1275) apparently already stated, e.g. as he said that species selection can utilize and maintain such "traits of species-individuals" [i.e. flexibility at the organismal level, such as generalised anatomy and wide ecological tolerance], but cannot "fashion" such traits, "that [=since they] (...) originate by conventional organismal selection".
Similarly, organismal selection can use and maintain but cannot fashion flexibility at the genetic level, since it originates by conventional genetic selection (see above the examples of transposons and junk DNA).
Perhaps there is no 'higher level' selection (in terms of differential reproductive output). Perhaps species evolve features that spread themselves though a population and as long as they do not effect differential reproductive output they establish themselves (such as in biased gene conversion). Gould etc made a lot of verbiage adaptations vs exaptations etc. but it seems to me just a lot of verbiage that ties one up in a tangled web of trying to distinguish between things that may not even exist.
John if I understood what you meant, if one accepts the possibility of species selection (e.g. re Fisher's objection about the limitations of low N), it is not a question of "reproductive output" at that level. Darwinian individuals at the species level are species, not organisms. The "reproductive rate" of species is the speciation rate; at the species level, "mutations" are organismal variants. The evolutionary dynamics are quite different (e.g. the role of structural constraints). I know it is not easy to read a lot of words (and Gould probably liked to write a bit too much), but I guess one could make an effort if it is worth doing it :-)
Gianluca, I've read Gould's work well enough. Its not the amount of words that it the problem, but the obfuscation that often went along with them.
I was not referring to species selection in any Darwinian sense or making any differential comparison with organisms. I guess what was lost in my text was the view that natural selection processes may not necessarily be involved in either the origin of speciation or adaptation (adaptation used in the sense of any recognized functional interaction between a structure and the environment). Gould, of course, saw natural selection as the only mode by which traits are established, as you quote "originate by conventional organismal selection.
John, you say that selection may not be involved in "functional interactions" between structure and environment. But if the interaction has no effect on fitness, what do you mean by "functional"?
At the gene hierarchical level, biased gene conversion (e.g. like concerted evolution), is a sort of molecular structural constraint driving evolution, not unlike exaptations at the organismal level. As we said, mutations or exaptations are not "random" in a statistical sense (i.e. they are always biased or constrained by specific mechanisms, e.g. at the molecular or developmental level), but they are random with respect to the future adaptive evolutionary directions, determined by the contingent selective regime. Of course if the bulk of evolutionary trajectories is determined/biased by non-random structures and mechanisms, then the role of selection would become much less important. But this is exactly what Gould was afraid of, when he catalysed the general attention upon exaptations.
I think Gould (2002) was actually worried that his attempt to revive the importance of structuralist theoretical themes (positive and negative constraints, at different levels of his hierarchical theoretical structure, i.e. including biased gene conversion at the gene level) would have been strongly rejected by adaptationists, and was a bit apologetic in this direction. However, though he still thought that selection is the strongest evolutionary force, he always said he was a pluralist, not an adaptationist, and that his contribution was actually a strong proposal of the importance of these very constraints (at different hierarchical levels) in the Modern Synthesis paradigm.
Yeah, I remember his appeal to 'pluralism' which I regard as an unintegrated approach (just a mixture of anything and everything)
Functional - just anything that one recognizes as a function. No implications about fitness.
One sees many absurdities in the literature about functions as the origin of structures (that awful intelligent design type of thinking). Classic example of bird feathers - supposed to be for (teleology) flight (adaption for flight) then of course in dinosaurs so they figure features have to be an adaptation for something else (catching bugs, keeping warm etc.). And on and on it goes. Crazy stuff.
Just to avoid running around in circles. My view is that there are two mechanisms that result in species specific genetic change - biological (molecular drive etc) and environmental (natural selection). Both may interact. Trying to create distinctions between exaptations and adaptations etc gets one involved in making distinctions between things that may be imagined, but not necessarily real.
Interesting that you suggest that Gould allowed potential peer pressure to modify his stance. My personal view is that he was more of a compiler than a synthesis. He certainly made sure never to mention the one work that gave an explicit and comprehensive non-Darwinian synthesis for the origin of adaptation (Croizat) even though he admitted, in writing, that he had read the works.
1. Pluralism in the MS just means acknowledging measured contributions of the main evolutionary forces: selection, drift, gene flow and mutation.
2. Either a structure has an effect on the fitness of the organism, then being either maladaptive or adaptive, or it hasn't, thus being non(ad)aptive. In the latter case, it just does not have any function (neutral).
3. Exaptations (shifts of function from either an adaptation or a nonaptation) are real stuff, not dreams. History is not a dream (though I think it is a form of imagination biased by reality :-) ). Gene duplications, pseudogenes and genetic shifts of functions are the results of historical events, not dreams. The same goes for the functional interpretation of structures such as feathers in modern birds, or in dinosaurs. Bird feathers are not "for" (teleology) flight. They are the end result of a historical process made of contingencies and interactions between biological systems and environments: the result is feathers+flight. Different disciplines such as molecular biology or paleontology just have different quantitative and qualitative methods.
4. Maybe I did not explain myself well in my previous message. I did not mean Gould allowed peer pressure to modify his views. I meant that he specifically aimed to contrast adaptationism by reinforcing structuralist themes, which seems quite the opposite of what you're saying.
Gianluca,
Thanks for the feedback. My comments below:
1. Pluralism in the MS just means acknowledging measured contributions of the main evolutionary forces: selection, drift, gene flow and mutation.
But there is no mention of mechanisms for the origin of mutation and their spread through populations without selection, drift, or gene flow.
2. Either a structure has an effect on the fitness of the organism, then being either maladaptive or adaptive, or it hasn't, thus being non(ad)aptive. In the latter case, it just does not have any function (neutral).
If you wish to define function that way.
3. Exaptations (shifts of function from either an adaptation or a nonaptation) are real stuff, not dreams.
Not so sure about that.
4. Gene duplications, pseudogenes and genetic shifts of functions are the results of historical events, not dreams.
Sure, but the dreaming comes in when people invoke ‘selective pressures’ or other such without any evidence.
5. Bird feathers are …the end result of a historical process made of contingencies and interactions between biological systems and environments:
true of anything (and anyway, a feature is really two feathers in one)
6. Gould ….. aimed to contrast adaptationism by reinforcing structuralist themes, which seems quite the opposite of what you're saying.
OK. Although I think he just further muddied the waters of Darwinian evolution. But that is just a personal opinion.
Cheers, John
p.s. interesting that Gould, like all other Darwinian theorists, always avoided discussion Darwin's 'laws of growth', a concept more in line with orthogenetic mechanisms than traditional Darwinism.
pps - obviously that should have read "a feather is really two feathers in one"
not sure about what you mean by this expression... two (or more, non-mutually exclusive) functions for the same structure, that evolved sequentially?
not sure to whom the question is directed about "the expression of two or more functions for the same structure, that evolved sequentially" as I did not use the expression and cannot find reference to it in the discussion.
John, I meant: what do you mean by "a feather is really two feathers in one" ?
PS interesting that your "true of anything" seems to philosophically imply that there is no historical sequence of causes and effects, that can be discriminated by temporally correlated events. It might be an interesting point: "things just happen together, not one after (due to) another", but of course the whole Darwinian paradigm, as any historical method, assumes this can be done.
- Badges
- Science topic
- Similar topics
- Instrumentation Engineering
- Instruments