Developmental Plasticity and Environmental Cues
A single water flea genotype can grow a helmet-shaped head, or grow a smooth round one. The difference is not a mutation but a memory of danger. If you raise Daphnia pulex alongside chemical cues from a predatory Chaoborus larva, the offspring sprout defensive neck spines and helmets. Raised without the cue, genetically identical siblings stay hornless. The genome did not change. What changed was which developmental program the genome ran. This was triggered by a chemical whisper from the environment. This is Phenotypic Plasticity — the capacity of a single genotype to produce different phenotypes depending on the conditions it develops in. It forces a correction to the naive view that genotype alone specifies phenotype. Developmental plasticity matters because it reframes the central dogma of development. Genes do not encode a fixed outcome. Instead, they encode a norm of reaction — a rule for translating environmental input into phenotypic output. This is not noise or error. It is often exquisitely adaptive, shaped by natural selection to track a variable world. Examples include seasonal color changes, caste determination in social insects, and sex determination in reptiles and fish. They also include migratory versus resident forms in locusts and aphids. Even human traits like adult height and disease susceptibility depend on childhood nutrition. Understanding plasticity requires the same rigor used for C. elegans and the Invariant Cell Lineage. Except that instead of an invariant, hard-wired program, we now ask how a conditional program is built. What senses the cue, what signal transduces it, and what developmental switch it throws. The stakes go beyond individual organisms. Plasticity is a major current in evolutionary theory. It can precede genetic change rather than merely follow it — a phenomenon called genetic assimilation. This was first demonstrated experimentally by Conrad Waddington. When a plastic response is repeatedly favored by selection, populations can evolve to produce the "induced" phenotype even without the inducing cue. This refines a plastic response into a fixed, canalized one. This connects directly to how Regeneration: Restoring Form After Loss and other environmentally sensitive developmental programs interact with evolution. It also sets the stage for Evolutionary Developmental Biology: The Evo-Devo Synthesis. There, plasticity is treated as a source of evolutionary novelty rather than a mere nuisance for geneticists. This article builds the concept from first principles. It covers what a reaction norm is and how it is measured. It explains the molecular and cellular machinery that senses environment and rewires development (polyphenism switches, hormonal titers, epigenetic marks). It presents the classic experiments that proved plasticity is real and heritable-in-its-capacity. It includes worked quantitative examples. Finally, it shows where these ideas are deployed today in medicine, agriculture, and conservation.
- Phenotypic plasticity: one genotype, multiple possible phenotypes, depending on the environment during development.
- A reaction norm is the quantitative map from environmental input to phenotypic output for a given genotype.
- Polyphenism is a special, discontinuous case of plasticity: two or more discrete morphs, not a smooth gradient.
- Environmental cues are sensed by receptors/sensory systems and relayed through hormonal or epigenetic switches (e.g., juvenile hormone, DNA methylation) that redirect gene expression during a sensitive window.
- Waddington's genetic assimilation experiments showed a plastic response can become fixed (canalized) under repeated selection — plasticity can precede and guide evolutionary change.
- Modern applications: predicting climate-driven sex ratio collapse in reptiles, understanding developmental origins of adult disease, breeding climate-resilient crops and locust-outbreak forecasting.
Development is not the readout of a fixed genetic blueprint but the execution of a conditional program — a reaction norm — in which environmental cues, sensed during defined critical windows, switch gene expression networks to produce alternative, sometimes discrete, phenotypes from one genotype.
The core idea
Phenotypic plasticity is formally defined as the ability of a single genotype to produce different phenotypes as a function of the environment it experiences during development (or, for some traits, throughout life). It is measured with a reaction norm: a function, or graph, relating an environmental variable (temperature, density, predator cue concentration, nutrition) on the x-axis to a phenotypic trait value on the y-axis, for one genotype. A flat reaction norm means the genotype is canalized — insensitive to that environmental range, buffered against perturbation (a concept formalized by Waddington as canalization, visualized by his famous "epigenetic landscape" metaphor). A sloped or curved reaction norm means the trait is plastic. Crucially, different genotypes can have different reaction norms for the same trait — this genotype-by-environment interaction (G×E) is exactly what selection acts on when plasticity itself evolves. A useful distinction separates continuous plasticity from polyphenism. Continuous plasticity produces a graded trait — fish grow more vertebrae in cold water, plants grow taller in shade — tracking the environmental variable smoothly. Polyphenism produces discrete, alternative morphs from one genotype with no intermediate forms in nature: locust solitary versus gregarious phases, aphid winged versus wingless morphs, Daphnia helmeted versus unhelmeted heads, temperature-dependent sex in many reptiles, and the queen/worker castes of honeybees. Polyphenisms are threshold phenomena: an underlying continuous physiological signal (a hormone titer, a nutritional index) is read out through a nonlinear, often bistable switch, so the organism commits to one of a small number of discrete developmental trajectories rather than hedging with an intermediate. Both forms of plasticity depend on the same three ingredients: a cue that reliably predicts future conditions, a sensor/signal-transduction system that detects the cue during a critical or sensitive window of development, and a developmental switch — usually a hormonal titer crossing a threshold, or a chromatin-level change — that redirects gene expression networks toward one alternative program or another. Because the switch acts upstream of the gene networks that build the trait, the same genotype can be wired to build either outcome; which one it builds is contingent on timing and dose of the cue. This is mechanistically continuous with regeneration biology, where positional information and signaling gradients likewise redeploy a single genome toward alternative structural outcomes depending on context.
Mechanism
Step 1 — reliable environmental cues. Selection favors plasticity only when the environment varies predictably within an organism's lifetime or across generations in a way an individual (or its parent) can detect before the consequences arrive. A chemical kairomone from a predator, day length (photoperiod) as a proxy for season, population density as a proxy for future competition, and incubation temperature as a proxy for later thermal environment are all such reliable, anticipatory cues. If the cue is unreliable, plastic responses are selected against because they produce a mismatched phenotype.
Step 2 — sensory and neuroendocrine detection. Cues are transduced by conventional sensory and biochemical machinery — chemoreceptors for kairomones, photoreceptors and circadian clock genes for photoperiod, thermoreceptors for temperature — feeding into central neuroendocrine centers. In insects, this typically converges on the brain–corpus allatum–prothoracic gland axis: neurosecretory cells integrate cues and adjust release of juvenile hormone (JH) and ecdysone (via prothoracicotropic hormone, PTTH). In vertebrates, analogous integration occurs via the hypothalamic–pituitary axes and steroid hormones (e.g., estrogen/testosterone ratios in temperature-dependent sex determination).
Step 3 — the developmental switch (threshold/bistability). The key feature that converts a graded internal signal into a discrete polyphenism is a nonlinear switch, often built from mutual antagonism or positive feedback between two transcriptional or signaling states, so that intermediate hormone levels are unstable and the system snaps to one of two attractor states. In aphids, crowding (density cue) is sensed and — via an as-yet only partially resolved endocrine/neuronal pathway — determines whether offspring develop wings; in the pea aphid, JH signaling has been shown experimentally to control the wing/wingless switch, since topical JH-analogue application induces winged morphs even at low density. In locusts, crowding triggers a serotonin surge that is necessary and sufficient (shown by injection experiments) to flip solitary-phase individuals toward the gregarious phase within hours to days, altering color, shape, and behavior. In honeybees, differential feeding of royal jelly to female larvae represses developmental pathways that would otherwise produce a worker, instead permitting ovary development and queen morphology; this operates partly through diet-dependent DNA methylation of genes such as Dnmt3-target loci, and silencing Dnmt3 by RNAi in worker-destined larvae is sufficient to shift many individuals toward queen-like development, directly linking an epigenetic writer to the caste switch.
Step 4 — gene regulatory network re-wiring. Downstream of the hormonal switch, transcription factor networks are differentially deployed. The same structural genes (cuticle proteins, patterning genes, growth-regulating genes) are present in both morphs; what differs is which combinatorial set is activated, when, and for how long, and at what growth rate imaginal/growth tissues proliferate. This is why plasticity does not require new genes — it requires new regulatory logic layered on existing genomic toolkits, the same principle central to Evolutionary Developmental Biology: The Evo-Devo Synthesis.
Step 5 — critical windows. The switch is only responsive during a defined developmental time window — for temperature-dependent sex determination in reptiles, this is the middle third of incubation (the thermosensitive period); outside that window, temperature has no effect on sex. This windowing means the same organism transitions from plastic to canalized for a given trait as development proceeds, echoing how C. elegans and the Invariant Cell Lineage shows some cell fates becoming irreversibly fixed once a critical developmental decision point passes.
Step 6 — epigenetic persistence. Some environmentally induced states persist because the switch event leaves a stable chromatin mark — DNA methylation, histone modification — that is clonally propagated through subsequent cell divisions without requiring the original cue to still be present. This explains why a phenotype triggered early (e.g., queen-fate in a bee larva, or an early-life nutritional signal in a mammal) can remain fixed for the rest of the organism's life even after the inducing signal disappears.
The evidence
Woltereck (1909) and the origin of the reaction norm concept. Richard Woltereck raised Daphnia clones (genetically uniform lineages, since Daphnia reproduce parthenogenetically) under varying food concentration and temperature and measured head/helmet shape and body size. Different clones showed different, characteristic curves relating environment to phenotype — Woltereck coined Reaktionsnorm (reaction norm) for this clone-specific function. This established, decades before molecular genetics, that "genotype" should be understood as specifying a norm of reaction, not a fixed trait, and that genetically identical individuals can look strikingly different depending on rearing conditions.
Modern Daphnia-*Chaoborus kairomone experiments.</strong> Building on Woltereck's system, later researchers (e.g., Christian Laforsch, Ralph Tollrian, and collaborators from the 1990s onward) showed that clonal Daphnia raised in water conditioned by predatory phantom-midge larvae (Chaoborus) develop pronounced helmets and neckteeth, while siblings from the same clone raised in predator-free water remain smooth-headed. Because the Daphnia* are genetically identical clones split across treatments, the only variable is the chemical cue in the water — a clean, controlled demonstration that a specific, isolable environmental signal (not genetic difference) causes the discrete morphological switch, and that the defense is adaptive (helmeted morphs suffer substantially lower predation from gape-limited predators in follow-up survival assays).
Waddington's genetic assimilation experiments (1942, 1953, 1956). Conrad Waddington asked whether an environmentally induced phenotype could become a heritable, unconditional one. He heat-shocked Drosophila melanogaster pupae (temperature stress, ~40 °C for a defined period during a sensitive developmental window), which induced a fraction of adults to eclose with a break in the wing crossveins ("crossveinless" phenocopy) — a phenotype normally produced only by a distinct genetic mutant, not by wild-type flies under normal conditions. Over successive generations, Waddington selected and bred only from the individuals that showed the induced crossveinless phenotype under heat shock. Within about 14 generations, a substantial fraction of flies from these selected lines showed the crossveinless phenotype even when no longer heat-shocked — the trait had become genetically assimilated: canalized into unconditional expression by selection acting on standing genetic variation in the population's threshold for responding to the stress. This is the single most important experimental proof that plasticity is not evolutionarily inert — a plastic, environmentally-triggered response can be converted by selection into a fixed trait, providing a concrete, non-Lamarckian mechanism (selection on genetic variants affecting the reaction norm's threshold, not inheritance of acquired characters) by which environmentally induced novelty can become a permanent feature of a lineage. Waddington repeated the logic with a different trait (the "bithorax" phenocopy induced by ether vapor), obtaining the same qualitative result, which strengthened the generality of the finding.
Nijhout and the hormonal basis of polyphenism (1990s–2000s). H. Frederik Nijhout's classic work on Precis coenia (buckeye butterfly) and other Lepidoptera used direct hormone manipulation — ligation experiments, hormone injections, and measurement of ecdysteroid/JH titers at defined larval stages — to show that seasonal wing-color polyphenism (a lighter "summer" form versus a darker, patterned "fall" form) is controlled by the timing of the ecdysteroid peak relative to a temperature-sensitive period in the pupa, not by different genes. Injecting ecdysteroid at the wrong time reproduces the alternative morph, directly demonstrating hormone titer as the proximate switch, which he generalized into an influential quantitative threshold model of polyphenism applicable across insect systems (locust phase change, aphid wing dimorphism, beetle horn dimorphism).
Temperature-dependent sex determination (TSD): incubation experiments in reptiles. In many turtles, all crocodilians, and some lizards, sex is not fixed by sex chromosomes at fertilization but determined by incubation temperature of the egg. Classic experiments (pioneered by Madeleine Charnier in Agama lizards, 1966, and extended by others across species) incubated eggs of a single clutch — controlling for genotype — at different constant temperatures and scored hatchling sex by gonad histology. In many turtle species, cooler incubation produces mostly males and warmer incubation produces mostly females (in some crocodilians the pattern is the reverse, with intermediate "male" windows), with a narrow pivotal temperature (often within about 1–2 °C) separating nearly all-male from nearly all-female clutches. Reciprocal transfer experiments — moving eggs between temperatures only during the thermosensitive period (roughly the middle third of incubation) — proved the switch acts in a defined developmental window, not throughout incubation.
Worked examples
Example 1 — Reading a reaction-norm/pivotal-temperature dataset in a turtle (quantitative reasoning on TSD). Suppose a researcher incubates eggs from a single clutch of a turtle species (genotype effectively controlled, since all eggs share a mother and similar paternity) at five constant temperatures and records percent female:
| Incubation temp (°C) | % female offspring |
|---|---|
| 25 | 4% |
| 27 | 20% |
| 28 | 52% |
| 29 | 84% |
| 31 | 98% |
The pivotal temperature (50% sex ratio) falls between 27 °C and 29 °C, close to 28 °C. This is a classic sigmoidal reaction norm: below ~26 °C the clutch is essentially all-male, above ~30 °C essentially all-female, and only a narrow ~2–3 °C window produces mixed sex ratios. The biological stakes of this steepness are large: a 1 °C rise in mean nest temperature, well within the range of projected climate warming, can shift a population from a healthy ~50:50 sex ratio toward >90% female, as has been documented in green sea turtle (Chelonia mydas) rookeries in the northern Great Barrier Reef, where recent cohorts have been found to be over 99% female in the warmest nesting beaches. Applying the logic of a reaction norm here converts a molecular/endocrine mechanism (temperature acting on aromatase expression, which converts testosterone to estrogen in the embryonic gonad during the thermosensitive period) directly into a population-level, conservation-relevant prediction: continued warming risks a demographic collapse from lack of males, independent of any change in the turtles' genotypes.
Example 2 — Genetic assimilation, worked as a selection problem. Model Waddington's crossveinless experiment abstractly. Suppose that within a Drosophila population, the threshold hormone/developmental sensitivity for producing a crossveinless wing under heat shock is controlled by multiple loci with additive effects, so that threshold value is normally distributed across the population (a standard quantitative-genetics assumption, consistent with what Waddington himself proposed). Let the environmental heat shock impose a fixed "stress level" S. Any fly whose threshold lies below S develops crossveinless wings; flies with a higher threshold do not, even though all experience the same heat shock.
- Generation 0: population mean threshold = μ₀, some fraction p₀ (a minority, say ~5–40% depending on the specific study/line) falls below S and shows the phenotype.
- The breeder selects only crossveinless-under-stress flies each generation and breeds them.
- Because threshold is heritable (additive genetic variance present), this is directional selection on a quantitative trait: the mean threshold of the breeding population is pulled downward each generation (a straightforward application of the breeder's equation, response = heritability × selection differential).
- After ~14 generations of such selection, mean threshold μ₁₄ has shifted so far below typical unstressed physiological conditions that a substantial fraction of individuals express the crossveinless phenotype without any heat shock at all — the trait is now canalized/assimilated.
The key conceptual result: no new mutation for "crossveinless" needed to arise during the experiment. Selection simply enriched for pre-existing genetic variants (in stress-response/threshold genes) that lowered the developmental threshold, until the phenotype no longer required the environmental trigger to be expressed. This is a rigorous, non-Lamarckian mechanism: the environment (heat shock) did not instruct heritable change directly; it revealed which genotypes to select, and ordinary Mendelian selection on standing variation did the rest.
Common misconceptions & traps
| Misconception | Correction |
|---|---|
| "Plasticity means the trait is not genetically controlled." | The capacity for plasticity (the shape of the reaction norm itself) is genetically encoded and heritable; selection can act on reaction norms just as on fixed traits. |
| "Genetic assimilation is Lamarckian inheritance of acquired characters." | It is ordinary selection on pre-existing genetic variation in developmental thresholds; the acquired phenotype itself is never transmitted through the germline as such — only genotypes that produce it more readily are enriched. |
| "Polyphenism and continuous plasticity are the same thing." | Polyphenism produces discrete, non-overlapping morphs via a threshold switch; continuous plasticity produces a graded trait tracking the environmental variable smoothly. |
| "A cue that changes phenotype is instructive/adaptive by definition." | Some environmentally induced phenotypes are pathological byproducts (teratogenesis, phenocopies) with no adaptive value; only cue-response pairs shaped by selection to track a reliable, anticipatory signal count as adaptive plasticity. |
| "Epigenetic marks from environment are automatically passed to offspring (transgenerational)." | Most environmentally induced epigenetic marks are reset during gametogenesis/early embryogenesis; documented transgenerational epigenetic inheritance in mammals is real but comparatively rare and mechanistically distinct from within-generation plasticity. |
| "Temperature-dependent sex determination means these species lack sex chromosomes at all." | TSD and genotypic sex determination (GSD) are not mutually exclusive across the tree of life; some species show mixed systems where genotype sets a baseline that temperature can override near a threshold. |
The scientist behind it
Conrad Hal Waddington trained originally as a geologist and paleontologist before turning to embryology in the late 1920s, drawn to Cambridge and to the experimental embryology emerging from the Spemann school in Germany. He became one of the great synthetic thinkers of twentieth-century biology, working simultaneously as embryologist, geneticist, and philosopher of biology, and he is the person who gave developmental biology two of its most durable metaphors: canalization, the buffering of development against perturbation, and the epigenetic landscape, a ball rolling down a branching valley system representing a cell's or organism's constrained developmental trajectory. He coined the modern usage of the word "epigenetics" itself, defining it as the study of the causal interactions between genes and their products that bring the phenotype into being — a broader and more mechanistic sense than the narrower molecular meaning ("heritable chromatin marks") the word acquired later. Waddington worked across an unusually wide range of organisms and problems — Drosophila development and genetics, chick and amphibian organizer transplantation experiments continuing the Spemann–Mangold tradition, and later evolutionary theory — while also serving as a prominent public intellectual, writing on the philosophy of science and organizing the influential "Theoretical Biology Club." His genetic assimilation experiments were conducted at the Institute of Animal Genetics in Edinburgh, where he built a research program explicitly designed to test whether development and evolution could be linked through selectable variation in reaction norms, anticipating by decades the evo-devo synthesis that would formalize these ideas using molecular tools.
Origins · 1 minConrad Hal Waddington and how this was discovered
Where this matters today
Conservation biology and climate change. Temperature-dependent sex determination makes marine turtles, many freshwater turtles, and crocodilians directly vulnerable to warming nest temperatures; conservation programs now shade or water nesting beaches, or relocate clutches to cooler microsites, specifically to counteract predicted feminization documented in populations such as northern Great Barrier Reef green turtles. Understanding the reaction norm (not just "warming is bad" but the actual pivotal temperature and window width) lets managers make quantitative, testable interventions rather than guesses.
Developmental origins of health and disease (DOHaD). In humans, prenatal and early-life nutritional and hormonal environments shape long-term physiology — a plasticity phenomenon formalized in the "thrifty phenotype" hypothesis (Hales and Barker), where fetal undernutrition programs metabolism toward energy conservation, which becomes maladaptive (raising risk of type 2 diabetes and cardiovascular disease) if the postnatal environment turns out calorie-rich rather than calorie-poor — a mismatch between the "predicted" and "actual" environment. This directly extends the reaction-norm logic: the same genotype, different early environment, different (and here, clinically consequential) adult phenotype.
Agriculture and pest forecasting. Locust phase polyphenism (solitary versus gregarious) underlies devastating agricultural swarms; understanding the density-dependent serotonergic switch allows forecasting models that predict when populations will tip toward the gregarious, swarming phase, informing early intervention (targeted pesticide application before mass aggregation) used by agencies such as the UN Food and Agriculture Organization's Desert Locust Information Service.
Crop breeding for plastic resilience. Rather than breeding a single "best" genotype, agricultural breeders increasingly select crop varieties with favorable reaction norms — flat (canalized) yield across variable rainfall, or plastic root-architecture responses that forage more efficiently under drought — an approach directly descended from the Woltereck/Waddington conceptual framework and essential for climate-resilient agriculture.
Biomedicine and regenerative contexts. Epigenetic switch mechanisms first characterized in polyphenism research (e.g., DNA methylation writers controlling caste fate in honeybees) inform how researchers think about cellular plasticity in mammalian regeneration and reprogramming — directly continuous with themes in Regeneration: Restoring Form After Loss, where differentiated cells can sometimes be induced back toward more plastic states by manipulating analogous chromatin-modifying machinery.
Ask the tutor: How does a bistable hormonal switch actually produce a sharp threshold instead of a graded response — what molecular feedback loop makes that happen?
Is phenotypic plasticity itself heritable, or just noise?
The capacity for plasticity — the shape and slope of the reaction norm — is a heritable trait in its own right, controlled by genes affecting sensory transduction, hormone signaling, and gene-regulatory thresholds. Different genotypes within a population can show different reaction norms for the same trait (a genotype-by-environment interaction), and this variation can respond to selection. So plasticity is not developmental noise; it is itself an evolvable trait, which is precisely why Waddington could select on it and shift where a population's threshold sat, producing genetic assimilation.
What is the difference between phenotypic plasticity and simple developmental noise or error?
Developmental noise is stochastic variation in outcome that is not systematically tied to any particular environmental variable and is typically buffered against by canalization. Plasticity, by contrast, is a systematic, reproducible mapping from a specific environmental cue to a specific phenotypic outcome, and — when adaptive — has been shaped by selection because the cue reliably predicts future conditions. A teratogen causing random malformations is noise/damage; a Daphnia helmet reliably appearing only in predator-cue water is plasticity.
Can plasticity itself evolve to disappear, becoming a fixed trait?
Yes — this is exactly what genetic assimilation demonstrates. If an environment becomes so constant that the cue is always present (or always absent), selection can favor genotypes that produce the "induced" phenotype unconditionally, since maintaining a costly sensing-and-switching apparatus for an environment that no longer varies offers no benefit. Over generations, the reaction norm flattens: the trait becomes canalized (fixed) rather than plastic, even though it started out as a purely environmentally triggered response.
How does temperature-dependent sex determination actually work at the molecular level?
During a species-specific thermosensitive window in mid-incubation, temperature influences the expression of genes in the embryonic gonad, notably the enzyme aromatase, which converts testosterone into estradiol. Temperatures that favor high aromatase expression push the bipotential gonad toward ovarian development (female), while temperatures favoring low aromatase expression allow testis development (male), in species with the "warm = female" pattern; the exact upstream thermosensor and full gene network remain active research areas, but experimental aromatase-inhibitor and estrogen-injection studies confirm this pathway is causally central.
Is human height a case of phenotypic plasticity?
Partly, yes. Adult height in humans has substantial heritability within a given environment, but the reaction norm for height is clearly plastic: secular increases in average height across the 20th century in many countries (several centimeters per generation in some cohorts) occurred far too fast to reflect genetic change and instead reflect improved childhood nutrition and reduced disease burden altering the same genotypes' growth trajectories. This is a textbook case of a continuously plastic trait responding to an environmental input (nutrition/health) rather than a discrete polyphenism.
Does phenotypic plasticity contradict or support Darwinian natural selection?
It fully supports it, once properly understood. Natural selection can act on the reaction norm (which genotypes produce which phenotype under which conditions) exactly as it acts on any other heritable trait; plasticity is neither an alternative to nor a violation of natural selection. Where plasticity gets historically entangled with controversy is genetic assimilation, which superficially resembles Lamarckism but is mechanistically pure Darwinian selection on standing variation in developmental thresholds — no inheritance of acquired characters is required or implied.
Why do biologists care about plasticity for evolutionary novelty (evo-devo)?
Because plasticity can generate a phenotypic variant before any genetic variant exists for it — an environmentally induced novel morphology can appear in a population, be exposed repeatedly to selection, and only afterward become genetically stabilized via genetic assimilation. This "plasticity-first" route to evolutionary novelty is an active research theme in evo-devo, since it offers a way for complex new phenotypes (like polyphenic morphs or novel skeletal structures) to arise and be tested by selection without first requiring a lucky new mutation, reframing how gene regulatory network innovation and organismal novelty are thought to originate.