🎓 How Organisms Rewrite Bodies Without Changing Genes

On phenotype plasticity, facultative neoteny, RNA editing, and metamorphosis, focused on locusts, maize, octopuses, dogs and holly trees respond to environments without "evolving" per se.

Apr 22, 2026

When I wrote about how locusts cause famines, I included a quick aside about octopuses and water temperature. It was just a tangent: I was trying to explain phenotypic plasticity and my brain jumped to cephalopods (because I’ve never seen an octopus that wasn’t cool). But upon reflection, I think I fudged some important details and was a bit imprecise. What octopuses do with their RNA is actually a different mechanism from what locusts and maize do with their genes, although both are methods organisms use to adapt to their environment on a physical level.

So here’s what I wish I’d said, expanded into its own deep dive into how creatures can change their bodies without “adapting” in the traditional, evolutionary sense.

What Phenotypic Plasticity Actually Is

The basic idea is that there’s one set of genes with multiple possible physical outcomes. In the original example, a grasshopper has the same DNA as the locust it later transforms into.

The genotype is an organism’s complete set of DNA instructions, and the phenotype is what the organism actually looks like and does. The genes don’t mutate and there isn’t any evolution-style selection pressure involved. The organism just reads its existing instructions differently depending on what’s going on around it.

This is pretty normal. There’s a study about how that phenotypic plasticity has been a significant factor in maize adaptation, alongside conventional selection. Humans display it too: we build more muscle with exercise, acquire tans after enough time exposed to the sun, and get taller if we have better nutrition (or skip our periods if the food situation is too bad or we exercise too much, like my grandma did when she was a young softball player). But some organisms take it to extremes.

I covered this in more detail earlier, but certain grasshopper species can transform into swarming, crop-devouring plagues when they get crowded together. They change color, breed faster, and start moving as a group after crowding causes physical contact, which triggers serotonin production that flips a behavioral switch.

This phenotypic plasticity and not metamorphosis because of the reversibility. Even fully switched-over locusts can revert to solitary grasshopper form if they are isolated. The behavioral changes start within hours, driven by dopamine instead of serotonin this time.

Then there’s maize. Scientists tested what happened when they grew teosinte (wild maize) under conditions mimicking the early Holocene climate, similar to when domestication actually occurred. The teosinte expressed more “domesticated” phenotypes — which is to say they grew smaller, with fewer branches — than the same plants grown in modern conditions. Apparently (forgive me, I asked Claude, this is confusing stuff) this suggests that “early agriculturalists were selecting for genetic mechanisms that cemented traits initially induced by a plastic response to the environment, a process called genetic assimilation.”

The gene under selection during domestication helps maize adjust flowering time depending on latitude and day length, although of course the farmers didn’t know how it worked on the micro level. Presumably they just noticed that some plants did better when moved to new fields.

In other words, maize may have started being useful because of where it was growing, and then humans locked in those traits through selective breeding so that it did the same thing no matter where you put it.

For an example that’s visible on a single tree, consider holly (which I hate, it’s the only type of tree I’ve ever cut down, pricked my feet soooo much as a kid barefoot in the backyard growing up). Carlos Herrera and Pilar Bazaga found in 2013 that holly trees produce prickly leaves where deer browse and smooth leaves above browsing height. Every leaf on the tree has identical DNA, but prickly leaves are significantly less methylated than smooth ones. The tree senses herbivory and adjusts its chemical markers to change how genes are expressed, producing physical defenses exactly where they’re needed. I mentioned this briefly in my piece on trees as infrastructure, which is a really great deep dive on how trees are criminally underutilized in speculative fiction… but anyway.

Octopuses: Something Different Entirely

When I mentioned octopuses in the locusts article, I said you see phenotypic plasticity “a lot when you look at how an octopus responds to different water temperatures.” That’s... mostly true, but it glosses over what makes cephalopods special.

What octopuses actually do (so far as I know — I’ve been wrong before tho 🙈) is RNA editing. This is a distinct mechanism from phenotypic plasticity.

In standard phenotypic plasticity (locusts, maize, holly), the DNA stays the same and the organism changes which genes it expresses, or how much of a given protein it produces. The instructions stay the same but the organism reads different parts of them. Like a box of legos that accommodates multiple builds depending on which directions packet you elect to start with (PS: Rebrickable is amazing for letting you re-spin different lego kits into different designs... caaaan you tell I have a six year old?)

But RNA editing is more like getting a not-so-artificial intelligence (heh) to actually go and re-write the directions. As I understand it, the process works like this: the cell copies a gene’s instructions into RNA (the working blueprint for making proteins), and then octopuses chemically swap out specific letters in that RNA before the protein gets built. One type of molecular letter gets changed to another, and the resulting protein ends up with a different building block than what the DNA originally called for. Think of it like a copy editor changing words in a manuscript after the author has finished writing but before it goes to print — the original stays intact, but what the reader actually gets is different.

Researchers found that when octopuses are moved from warm water (22°C) to cold water (13°C), they ramp up editing at over 13,000 protein-altering RNA sites — about a third of all their recoding sites — and the changes reach a steady state within days. These edits change proteins involved in neural signaling, membrane function, and calcium-dependent processes, allowing the nervous system to work properly at temperatures that would otherwise impair it. It really is a bit like an AI editing its own config files on the fly to handle a new operating environment (but with careful git ‘backups’ oh my goodness if you let AI edit things on your local files you need to have a good 3-2-1 backup process.)

Joshua Rosenthal, one of the lead researchers, said “the idea the environment can influence that genetic information, as we’ve shown in cephalopods, is a new concept.” Turns out an arctic octopus and a tropical octopus can share virtually the same DNA while producing functionally different neural proteins.

Tl;dr: RNA editing is faster and more reversible than most forms of phenotypic plasticity, and it targets specific proteins rather than whole developmental pathways.

Everything is Tradeoffs

There’s a cost to all this molecular flexibility, though. Coleoid cephalopods (octopuses, squid, cuttlefish) are trading transcription plasticity at the expense of genome evolution. The genomic regions around RNA editing sites are highly conserved, which means they can’t mutate much without breaking the editing machinery. So while cephalopods gain the ability to rapidly adjust their proteins, they lose the ability to evolve those same regions through conventional mutation. I wrote relating AI hallucinations, medieval copywriting errors, and the impacts of radiation on the human genome before, but basically (lol) none of the normal mutations actually “stick.”

Dogs and other domesticated creatures (like wheat!) are sort of the opposite. Domestication bottlenecks reduced effective population size, which weakened purifying selection, and mildly bad mutations that would have been purged from larger wolf populations accumulated instead. And when breeders dragged beneficial alleles to fixation for traits like size or coat type, they dragged bad nearby variants along for the ride — which is why purebred dogs have such high rates of breed-specific genetic diseases. In both cases, intense selection on one function constrains what can happen in nearby genomic regions. For cephalopods, that means function get preserved but dogs end up dragging along “harmful” passengers. The neutral case is probably something red hair and freckles and green eyes often being paired.

At the fastest end, signaling-based phenotypic plasticity (like the locust’s serotonin switch) can reshape an organism in hours to weeks and is usually reversible. Epigenetic plasticity, like holly’s methylation-driven leaf changes, operates over days to seasons and is semi-reversible. RNA editing, the octopus specialty, works in hours to days and is rapidly reversible. And at the slow end, genetic assimilation locks in plastic traits through selection over generations, as happened with domesticated maize.

There are different tools for different timescales and different kinds of environmental pressure, and organisms often use more than one.

& Then We Have Axolotls

I originally wanted to include metamorphosis as a point of comparison, not least of which because a buddy of mine bred them and a they’re just neat animals. But they’re another one of what I like to call “inversion cases” because axolotls reach sexual maturity without ever undergoing the metamorphosis that other salamanders go through. The term for this is “neotany.” They reach reproductive maturity around 12-18 months, and do cute little courtship dances and all even though the rest of their body never “grows up.”

Metamorphosis and sexual maturation are controlled by separate hormonal axes: the thyroid axis (which triggers metamorphosis) is what’s broken in axolotls, but the gonadal axis works fine. So they hit puberty without ever leaving childhood in every other respect. They keep their gills, their fins, their larval body plan, and probably other stuff I’m forgetting.

You can force an axolotl to metamorphose by injecting it with thyroxine, whereupon they absorb their gills, grow eyelids and a tongue, and become terrestrial salamanders.

Contrast that with other salamander species that can go either way. Biologists call this “facultative paedomorphosis,” meaning the animal has the option of staying larval or metamorphosing depending on conditions. In mountain environments, some salamanders stay in larval form because activity in the hypothalamo-pituitary-thyroid axis stays low — the hormonal cascade that would trigger metamorphosis never fires. Whether that’s because of iodine scarcity, temperature, food availability, or some combination isn’t fully settled. Tiger salamander larvae are also known to develop cannibalistic morphs under crowded conditions, which is its own fascinating example of phenotypic plasticity, though I could not find any evidence of Wikipedia’s claim1 that cannibalism provides enough iodine to trigger metamorphosis.

For axolotls, genetic modification to the thyroid resulted in permanent retention of juvenile traits. Domestic dogs do something similar, retaining puppy behaviors like barking, whining, and licking your chin to try and get food... all of which wolves outgrow. Behavioral neoteny in dogs is driven by selective breeding rather than a single hormonal pathway, but they both end up keeping juvenile features into adulthood.

I once wrote a story called the Magic of Marsh Protection, where the snappers trigger metamorphosis by eating enough — facultative metamorphosis ended up being a plot mechanic because the protagonist needed to kill the baby snapper before it got big enough to undergo what is basically a Pokemon-style “evolution” where it’s a much bigger problem for the locals. Pokemon calling this metamorphosis “evolution” is probably the worst thing about the game (my kid is gonna end up so confused when he hits that year in Science class...) although otherwise I think that Pokemon Go is an awesome game, as I explained back in November.

Yes, And?

“Adapt or die” implies an evolutionary process that is generational and permanent. But biology has developed multiple ways to change on the fly, each with different costs and benefits.

Facultative neoteny waits on resources before permitting metamorphosis into an adult state capable of breeding. RNA editing sacrifices long-term evolutionary flexibility for short-term protein flexibility. Phenotypic plasticity trades optimization for the ability to cope with variable environments.

Which is to say, oops, because maize breeders who reduced genotype-by-environment interaction through selection may have inadvertently limited the crop’s ability to handle novel environments — exactly the kind of flexibility it might need as the climate changes, globalization allows pests to spread into all sorts of nooks and crannies around the world, and cultural preferences spread and consolidate. I wrote about how agricultural crises cascade into societal collapse like maize mosaic disease and the Maya, wheat rust and Rome, chestnut blight in Appalachia… and the common thread is monoculture vulnerability.

Understanding when to breed for plasticity versus robustness is one of the central challenges for future food security, which is either promising (they know about the problem!) or terrifying (if it’s hard, maybe we can’t solve it in time to prevent the next potential famine). Which way do you lean?

Manuscriptions

Discussion about this post

Ready for more?