Today, I gave my final lecture in developmental biology this term. We have one more class session which will be a final discussion, but I’m done yapping at them. Since I can’t possibly teach them everything, I offered some suggestions on what to read next, if they’re really interested in developmental biology. They’ve gotten the fundamentals of the dominant way of looking at development now, that good ol’ molecular genetics centered modern field of evo-devo, but I specifically wanted to suggest a few titles to shake them up a little bit and start thinking differently.
-
For the student who is interested in the field, but doesn’t feel that development is necessarily their discipline, I recommended Richard Lewontin’s The Triple Helix: Gene, Organism, and Environment(amzn/b&n/abe/pwll). It’s short, it’s easy, and it’s a good counterweight to the usual gene-happy approach we see in developmental biology.
-
Since we are a liberal arts university, and we value a philosophical approach in addition to the usual bluntly pragmatic tactics we follow in the sciences, I also recommended one work of philosophy: The Ontogeny of Information: Developmental Systems and Evolution(amzn/b&n/abe/pwll), by Susan Oyama. That one is not an easy read, except maybe to the more academically minded. I mentioned that Developmental Systems Theory does not have the powerful research program that is making evo-devo so successful, but it’s still a usefully different way of thinking about the world.
-
If any of my students wanted to go on to grad school in developmental biology, and hoped to make it a profession, I had to tell them that they are required to read D’Arcy Wentworth Thompson’s On Growth and Form(amzn/b&n/abe/pwll). It’s old, it’s a little bit weird, but it’s still a major touchstone in the discipline.
-
Lastly, I told them that there was one more book they had to read if they wanted to consider a career in development: Developmental Plasticity and Evolution(amzn/b&n/abe/pwll), by Mary Jane West-Eberhard. If I were a young graduate student in the field right now, I think I could just open that book to a random page and find an interesting and challenging research problem right there. I might have to flip through a few dozen pages before I found one that wasn’t impossibly hard, but hey, it’s one of those books that fills you in on the array of issues that people are worrying over at the edge of the science.
I don’t think any of these would be a good foundation for an undergraduate course (either Thompson or West-Eberhard or Oyama would probably have a lethal effect on the brain of any unprepared student trying to plow through them), but they’d be great mind-stretchers for any student planning to move on.
So all my lecturing is done for the term, and all that’s left are monstrous piles of grading that will grow ominously in the next week and a half even as I struggle to keep up, and then I can try to polish it all off by Cephalopodmas.
When I finished my undergrad genetics and devo courses (back in the dark ages of the late ’80s), a professor gave me “The Evolution of Individuality” by Leo Buss. Not an easy read – it assumes a reasonable familiarity with a lot of descriptive embryology, for example – but at the time it inspired me to think a lot more deeply about the evolution of developmental pathways, and was among the more important of the things that changed my plans and diverted me into grad school.
Brother, I feel your pain. I’m in finals week right now myself and am looking at a mountain of exams from my Ecology and Evolution course as well as a minor foothill of lab practical exams from Vertebrate Natural History. But I raise my glass to you, from one educator to another. You have my admiration and respect as someone who is not only willing to take your science to the people but to take it into the classroom, as well. As my Aussie friends like to say, good on ya.
just wondering; is Gilbert’s devo text still popular for undergrad devo courses these days?
Ahhh, PZ, let me tell you about the joys of semi-retirement (from academia, at least). Teach just one course with enough prereqs to keep enrollment down; have all upperclass majors in the class (plus one stray faculty member sitting in and doing all the assigned work); give no examinations (lab projects to grade, but that’s kind of fun); and one three-hour class meeting per week. :)
The downside, of course, is that I’m doing it for no pay. Hm. What’s wrong with this picture? Nothing, really.
It is interesting that most of those authors (except old man Thompson) would agree that the fundamental evolutionary “core” is selection for inheritable traits. Gould did too. He does not even talk on subjects like genetic assimilation in the fat book (though GA is now a hot topic at the frontlines of evolutionary theory).
Should we therefore go to those who focus on the core? Hamilton, Wilson, Maynard Smith…brainy people, lots of theory.
I say no…because I think we can find a better, more realistic “core” description of evolution.
“just wondering; is Gilbert’s devo text still popular for undergrad devo courses these days?”
Ichthyic: I used it for my devo course this last spring, so maybe. I thought it was an excellent text, with lots and lots of visuals.:)
I use Wolpert, but Gilbert is definitely the text. I’ve found Wolpert to be a little lighter, which is desirable since I also use some other supplemental texts. If I were to use just one book, though, it would be Gilbert.
One way to make Q&A periods more fun and more involving for the class is to ask other students to answer the questions, either by asking for a volunteer or picking a victim.
Congratulations on making it through another term.
If I were to use just one book, though, it would be Gilbert.
thanks. it gets confusing sometimes prowling about looking for acceptable texts.
Gilbert was the one they were using at UCB back in the early 90’s when I was there, and I see it has stood the test of time.
I too, found it an excellent text (and have been referring folks with devo questions to it for years now), but since I don’t follow dev bio as much as i do behavioral ecology, I wasn’t sure if it was still considered the “defacto” standard or not.
re: Thompson. I think Gould almost nailed it when he called it “an unusable masterpiece.” It’s interesting that developmental biologists all love this book (I do too), yet it has nothing to say about anything developmental biologists have done for the last 30 years.
Thompsons completely dismissed genetics in favor of biomechanical processes that drive morphology. Meanwhile molecular genetics completely dismissed mechanical processes of morphogenesis (or decided that morphogenesis meant gene expression patterns) in favor of genes and “gene action,” whatever that meant.
It’s a fantastic time to be a developmental biologist… not just for evo devo coming to life, but cell biology has progressed to the point that Thomas Morgan and D’Arcy Thompson would have a lot to talk about if they were here today.
Looks like you’ve suffered the same treatment as talkorigins: there are a lot of gunk links at the bottom of this page.
Sorry, I take that back. The rest of the page displayed properly now, and it’s just spam.
Developmental Plasticity and Evolution by Mary Jane West-Eberhard lists much good and interesting organismal work that was outside the view of molecular and developmental people, but her own contribution consists of fuzzy concepts lacking clear thinking. Just try to figure out what she actually means to say in this hugely overblown book.
de Jong, maybe West-Eberhard comes off as an incoherent and ambiguous critique because she still believes there must be a “fundamentally genetic”, darwinian core. But in truth, it is quite difficult to have such a core, and at the same time take epigenetic plasticity seriously.
And epigenetic plasticity certainly deserves to be taken seriously…
didn’t Thompson begin all the grid comparisons of morphometric geometry, landmarks and such? It is a big mistake, therefore to say his contribution is “useless”. Molecular developmental biologists don’t know much about this, but paleontologists are familiar with these applications. However nowadays morphometrics is used along with QTL analysis and other neat stuff..
Vargas, Would epigenetic plasticity be the same as phenotypic plasticity? If yes, classical genetics works. If no, statistical genetics will still work. Unless epigenetic plasticity implies adaptation is possible without any previous selection? Then you have to specify the mechanism. Phenotypic plasticity gives adaptation over some environments as a consequence of previous selection on plasticity.
Vargas, Would epigenetic plasticity be the same as phenotypic plasticity? If yes, classical genetics works. If no, statistical genetics will still work. Unless epigenetic plasticity implies adaptation is possible without any previous selection? Then you have to specify the mechanism. Phenotypic plasticity gives adaptation over some environments as a consequence of previous selection on plasticity.
Selection will have always occurred previously, but the point is, can previous evolution just be attributed to gene selection? If at one point I can see that a new adaptation occurs by phenotypic plasticity, without a new mutation nor selection, isn’t it a bit tautological to assume this is only possible because of previous selection? Could similar steps without mutation and selection also been of importance in that previous evolution? We must be careful not to conflate “previous evolution” with “the result exclusively of gene selection”.
Have you ever seen a new adaptation by phenotypic plasticity, without a new mutation or selection, that is not an extension of previous phenotypic plasticity? Simplifying, if you have phenotypic plasticity along a known reaction norm from temperature 15°C to 20°C, and then put animals at 25°C, do you get the extrapolation of the previous reaction norm as phenotype? (Is that adaptive? How to show that?) Previous selection between 15°C and 20°C can be extrapolated if the phenotype is physiologically ‘similar’. A new adaptation by phenotypic plasticity, without a new mutation or selection, requires adaptive phenotypic plasticity to exist.
As to the ‘gene selection’, I care little whether the actual mechanism is by direct genetic influence as in allozyme polymorphism or by methylation or any epigenetic mechanism. That is not the point. The point is whether plasticity is ‘created ex nihilo’ by unknown mystic organismal properties. West-Eberhard never thought that through, and hides behind long words and convoluted writing.
As you have framed it “The point is whether plasticity is ‘created ex nihilo’ by unknown mystic organismal properties”. Not at all. Plasticity is always allowed by preexisiting organismal structure (including its genes) but not only that, there must also be environmental change, and THEN plasticity arises not as a mere sum of both but as a specific kind of interaction. This is more realistic than assuming the whole response is preformed as if “encapsulated” in previous genetic composition.
No one is proposing creation of phenotypic plasticity “ex nihilo” by “mystic” properties. We agree that previous conditions are part of allowing this capacity. But I repeat: “previous conditions” is not equal to “the result of gene selection”. That is precisely what we are debating: How non-genetic change can affect the course of evolution. If it can, those previous conditions can be the result of more than just gene selection.
We must be careful not to conflate “previous evolution” with “the result exclusively of gene selection”.
…and Vargas must be careful not to overgeneralize exceptions as the rule.
there is no “gene selection”; you ARE conflating misunderstandings of developmental plasticity and genetic expression, with selection on the resulting phenotypes.
I mean by “gene selection” when the frequency of a gene in a population increases through natural selection. What were you thinking?
then why not say that, rather than creating a term which sounds more like you are saying selection acts directly on genes?
Isn’t what you are really trying to address called frequency dependent selection?
Vargas writes:
“Plasticity is always allowed by preexisting organismal structure (including its genes) but not only that, there must also be environmental change, and THEN plasticity arises not as a mere sum of both but as a specific kind of interaction. This is more realistic than assuming the whole response is preformed as if “encapsulated” in previous genetic composition.”
These are the sort of statements I object to as the words don’t seem to mean anything specific to get a hold on. So, there is preexisting organismal structure (including genes) and there is a variable environment. What does “plasticity arises not as the mere sum of both but as a specific kind of interaction” mean? Does this add anything to our understanding of phenotypic plasticity? It will depend on the phenotypic trait studied what the actual mechanisms underlying phenotypic plasticity are, but does this statement reflect a study of mechanisms?
Why is it more realistic to assume that plasticity arises as a specific kind of interaction between organism and environment rather than regard it as an existing potential property of the organism? Why is this more realistic than “assuming the whole response is preformed as if “encapsulated” in previous genetic composition? I still read the statements as saying plasticity arises de novo. Plasticity is itself inherited: there is genetic variation for phenotypic plasticity in traits. What you have to find is an example of a new adaptation (that) occurs by phenotypic plasticity, without a new mutation nor selection and is not the extrapolation of the existing reaction norm.
Phenotypic selection on any quantitative trait – and phenotypic plasticity is a quantitative trait – will lead to changes in allele frequencies of some alleles at some loci, as long as there is variation at some genes that eventually participate in the development of the phenotypic trait. A plastic response can be selected on, including its shape.
I mean that the interactions between organism (including its genetic composition) and the environment usually preclude the response from being a simple sum of each. This is well-known. For instance, a certain response may require for a gene, and for a specific environmental condition. We may think both add up in the response, but if we remove only one, we can have no response, rather than the contribution of the remaining part. They must find each other and interact; the response emerges from that interaction.
This, of course, has a truly “de novo” ( “emergent” or “epi”) aspect to it, even though previous existence of both parts are required and make it possible. But hey!! This is no reason to despair, or support fallacies such as it MUST have “all already been there”, either in the genes alone, or in the environment alone. No reason to run to a guru. In fact, non-commutativity and complex interactions are more the rule than the exception in biology.
Icthyic, what do you mean by selection “acts directly” on genes? A transposable element with high replication? What about hypothetical “altruist genes” that spread despite leading the organism to sacrifice?
If a gene is selected through its direct effect on the phenotype, that doesn’t count? Well, that is what I meant (and is probably the most accepted mechanism by which genes are selected). I certainly did not mean frequency dependent selection.
Arguments of equivalence between the genotype and phenotype are consistent with the defense of a fundamentally genetic core for evolutionary science…
sorry, but there is no way from what you just wrote that I could make sense of what your point is.
I give up.
Brain overload? Let me edit
What did you mean when you said selection “acts directly” on genes? just give us an example. I hope you don’t overexert youself
Vargas, I can think of an environment without any visible multi-cellular organism but not of any organism without an environment, and genes without an organism are food in the form of DNA. If “a certain response may require for a gene, and for a specific environmental condition. We may think both add up in the response, but if we remove only one, we can have no response, rather than the contribution of the remaining part. They must find each other and interact; the response emerges from that interaction. you are just giving a complicated description of the phenotype. Nothing about “emergent” or “epi” increases our understanding of phenotype. And a genotype is selected through its effect on the phenotype, and genes & alleles are selected only as part of the genotypes they find themselves in.
However, in “the interactions between organism (including its genetic composition) and the environment usually preclude the response from being a simple sum of each” you seem to get confused between mechanism and model. I’ll take that the qualifier “usually” implies that you are accepting that in some cases phenotypic expression is independent of the environment. Hopi Hoekstra’s work on the rock pocket mouse shows simple Mendelian dark = dominant / light = recessive coat color in the Pinacate population. (PNAS 2003). There, there is a perfect association between genotype and phenotype.
However, in classical quantitative genetics the model P=G+E does not assert anything about the mechanism. What it intends to say is that the genotypic value is defined as the mean phenotypic value. So, take clone 1 of Daphnia, measure helmet size of many individuals in a defined environment: mean helmet size becomes genotypic value for that clone in that environment. Repeat over other environments and you’ve got a genotypic value as a function of some environmental variable. This is a genetic function, and another clone will have another function that gives the genotypic value as a function of the environment. The mechanism of getting those reaction norms might be any complex set of developmental interactions, but it is possible to write the function ‘genotypic value’ as a sum, that is, as a polynomial.
As to “assuming the whole response is preformed as if “encapsulated” in previous genetic composition”, I assume that in a Drosophila melanogaster population that is selected in nature between say 15°C and 22.5°C (from a temperate climatic region) and shows a mean wing length over this temperature range that is linear with the environment (y = 2.37 – 0.0413x, y wing length, 2.4 in mm, -0.04 in mm per °C, x °C), then I expect wing length at 27.5°C to be on that same line, that is, the wing to be 1.23 mm, as 27.5°C is yet stressingly high. That is by previous selection, an extrapolation based on the reaction norm for the genotypic mean of that population. It is genetic in the sense than other populations show other reaction norms, and crosses can be made. If one wants the reaction norm can be selected on. But wing length will not be 5 cm, even if that would be adaptive.
My point is that phenotypic plasticity allows this type of extrapolation – but not a novel wing length. If you want to get rid of a “fundamentally genetic”, darwinian core, and if you can see that a new adaptation occurs by phenotypic plasticity, without a new mutation nor selection you had better come up with an example. By the way, where does this distaste for “fundamentally genetic” come from?
“Vargas, I can think of an environment without any visible multi-cellular organism but not of any organism without an environment”
Precisely. A specific environment is a constitutive part of any organism.
“you are just giving a complicated description of the phenotype”.
I gave an example of non- additivity of genotype and environment. It is part of understanding the relation of genotype to phenotype.
“Nothing about “emergent” or “epi” increases our understanding of phenotype”
Emergent properties in a system are those that do not arise as a simple commutative sum of their components. According to Mayr and Gould, biology is full of emergent properties. The developmental concept of epigenesis also points to how interactions emerge within the embryo. I think the phenotype cannot be understood without delving into the notion of emerging properties.
“Hopi Hoekstra’s work on the rock pocket mouse shows simple Mendelian dark = dominant / light = recessive coat color in the Pinacate population. (PNAS 2003). There, there is a perfect association between genotype and phenotype. _However, in classical quantitative genetics the model P=G+E does not assert anything about the mechanism. What it intends to say is that the genotypic value is defined as the mean phenotypic value. So, take clone 1 of Daphnia, measure helmet size of many individuals in a defined environment: mean helmet size becomes genotypic value for that clone in that environment. Repeat over other environments and you’ve got a genotypic value as a function of some environmental variable. This is a genetic function, and another clone will have another function that gives the genotypic value as a function of the environment. The mechanism of getting those reaction norms might be any complex set of developmental interactions, but it is possible to write the function ‘genotypic value’ as a sum, that is, as a polynomial.”
I understand that different clones (genotypes) can have different reaction norms but I don’t see how does it help to call the mean size of a phenotypic trait in a specific environment a “genotypic value”. If they are the “dots” in a reaction norm, that includes the influence of the environment. If mean trait size were not also partly an “environmental value”, we would have no need to talk of reaction norms at all (such as the case of Hoekstra).
Now, you can write down the reaction norm as a polynomial and say this is because of mathematical convenience, so we do not mind that development in fact is full of complex, non-additive interactions. That is, the model only provides a mathematical description and has no expectations of corresponding to actual mechanisms of development. The problem arises when some terms of the mathematical description are given very concrete names, as if having a precise biological meaning.
“As to “assuming the whole response is preformed as if “encapsulated” in previous genetic composition”, I assume that in a Drosophila melanogaster population that is selected in nature between say 15°C and 22.5°C (from a temperate climatic region) and shows a mean wing length over this temperature range that is linear with the environment (y = 2.37 – 0.0413x, y wing length, 2.4 in mm, -0.04 in mm per °C, x °C), then I expect wing length at 27.5°C to be on that same line, that is, the wing to be 1.23 mm, as 27.5°C is yet stressingly high. That is by previous selection, an extrapolation based on the reaction norm for the genotypic mean of that population.
But wing length will not be 5 cm, even if that would be adaptive.
My point is that phenotypic plasticity allows this type of extrapolation – but not a novel wing length.”
Think about this: no matter how large the genetic potential for developing a large wing, at low temperatures no new maximum wing size may be expressed.
Environmental conditions can certainly allow or forbid a novel wing length.
“If one wants the reaction norm can be selected on”
Yes, but to do that, you know what step is first: environmental exposure.
“If you want to get rid of a “fundamentally genetic”, darwinian core, and if you can see that a new adaptation occurs by phenotypic plasticity, without a new mutation nor selection you had better come up with an example. By the way, where does this distaste for “fundamentally genetic” come from?”
Very well, a tennis player develops a larger right arm. We can say this is an “adaptation”. This trait originates in a specific behavioral interaction. Not as the result of any specific mutation. Not the result of generations of selection for increasingly larger right am size.
MJ West Eberhardt says there must be a “fundamentally genetic” theory of evolution. However, I disagree, I consider the focus should be placed on the phenotype