Ecology & Evolution — Populations, Communities, Ecosystems, Selection, Speciation

Ecology asks how organisms distribute themselves and interact in space and time; evolution asks how their lineages change across generations. The two fields share a common explanatory currency — differential reproduction in a particular environment — and they have fused into a single research programme since the Modern Synthesis. This note traces the hierarchy from individual organism up to biosphere, then runs the same hierarchy through evolutionary time: alleles, populations, species, clades, biota. Worked equations, named laws, the people who established them, and the canonical experiments are included so a downstream agent can reason about ecosystems and lineages without re-derivation.

1. Levels of organization

Biology organizes its objects of study in a strict spatial-and-causal hierarchy. Each level has emergent properties that cannot be reduced cleanly to the level below, which is why ecology and evolution are separate disciplines from molecular biology even though they share atoms and genes.

• Individual organism — one autonomous unit of metabolism, development, behaviour, and reproduction. Physiological ecology lives here: tolerance limits, energy budgets, thermoregulation, water balance.
• Population — all individuals of one species occupying a defined area at a defined time. The unit of evolutionary change in classical theory: allele frequencies shift in populations, not in individuals.
• Community — assemblage of all populations (often only the focal taxonomic subset, e.g. “plant community”, “fish community”) sharing a habitat. Interactions are the focus: competition, predation, mutualism.
• Ecosystem — the community plus its abiotic environment, treated as a flow-of-matter-and-energy system. Tansley introduced the term in 1935; H. T. Odum and Eugene Odum operationalized it through the 1950s with energy-flow diagrams.
• Landscape — a mosaic of patches of different ecosystems, large enough to contain multiple communities and the corridors / barriers between them.
• Biome — recurrent ecosystem types defined by climate and dominant life form (tropical rainforest, temperate grassland, boreal forest / taiga, tundra, desert, savanna, coral reef, pelagic open ocean, benthic, intertidal).
• Biosphere — the global sum of all biomes, the thin film of Earth that contains life, roughly from the Mariana Trench at 11 km depth to airborne microbes at ~80 km altitude.

The hierarchy is operationally important: a population biologist cannot answer ecosystem questions, and an ecosystem ecologist cannot answer individual-physiology questions, with the same data.

2. Population ecology

2.1 Density, distribution, age structure

Population density is N/area or N/volume. Methods of estimation include direct census (rare; only for large conspicuous organisms or trapping-saturated systems), quadrat sampling for sessile organisms (plants, barnacles), and mark-recapture for mobile ones. The Lincoln–Petersen estimator for closed populations is N̂ = (M·C)/R where M is the number marked on the first visit, C is the total caught on the second visit, and R is the number of those that were already marked. The Schnabel and Jolly–Seber extensions handle multiple sampling sessions and open populations respectively.

Dispersion patterns within a population fall into three idealized categories:

• Uniform — individuals more evenly spaced than chance (creosote bush Larrea tridentata via allelopathy; nesting seabirds via territoriality).
• Random — Poisson-distributed; rare in nature, requires no inter-individual interaction and no environmental heterogeneity.
• Clumped — most common pattern; driven by patchy resources, social grouping (herds, schools, flocks), or limited dispersal (clonal plants, fungal mycelium).

The variance-to-mean ratio of counts per quadrat diagnoses dispersion: ≈1 random, <1 uniform, >1 clumped.

Age structure (or stage structure) is critical for predicting future trajectory. A life table tabulates lx (survivorship to age x) and mx (fecundity at age x). Net reproductive rate R0 = Σ lx·mx, and when R0 = 1 a population is exactly replacing itself. Generation time T ≈ Σ x·lx·mx / R0, and the intrinsic rate of increase r ≈ ln(R0)/T for slow-growing populations (Lotka’s approximation).

Survivorship curves plot ln(lx) vs age and partition life-histories:

• Type I — low mortality through most of life, sharp drop at senescence (humans in industrialized countries, large mammals, oak trees).
• Type II — constant mortality risk per unit time, straight line on a log plot (many birds, lizards, rodents).
• Type III — high early mortality, low rate among survivors (oysters with millions of pelagic larvae, oak acorns, fish eggs, r-strategists generally).

2.2 Growth models

Exponential growth dN/dt = rN integrates to N(t) = N0 e^{rt}, with r the per-capita rate of increase in units of inverse time. r = b − d where b is per-capita birth rate and d per-capita death rate. Doubling time Td = ln(2)/r ≈ 0.693/r. Exponential growth is unrealistic over long horizons because it ignores resource limits, but it describes real systems over short windows: bacterial cultures in fresh medium, invasive species after introduction (cane toads Rhinella marina in Australia after 1935, zebra mussels Dreissena polymorpha in the Great Lakes after ~1988), epidemic spread in a naïve host population, human population from roughly 1700 to 1970.

Logistic growth introduces a carrying capacity K: dN/dt = rN(1 − N/K). The integrated form is N(t) = K / (1 + ((K − N0)/N0) e^{−rt}). Growth is fastest at N = K/2 (the maximum sustainable yield in fisheries theory under the Schaefer model). Pearl and Reed fit this to U.S. population data in 1920; Gause demonstrated it with Paramecium cultures in 1934.

Real populations rarely follow logistic curves cleanly. Allee effects (positive density dependence at low N — pollination limitation in plants, mate-finding in animals, group defense) cause sub-threshold populations to crash to extinction even when r at carrying capacity is positive. Discrete-generation models can show overshoot, damped oscillation, limit cycles, and deterministic chaos depending on r (the Ricker map N_{t+1} = N_t exp(r(1 − N_t/K)) and the logistic map Nt+1 = rNt(1 − Nt) of May 1976 are canonical demonstrations).

2.3 Life histories: r/K selection and beyond

MacArthur and Wilson formalized the r/K dichotomy in The Theory of Island Biogeography (1967). r-selected organisms maximize population growth rate: small body size, early maturity, many small offspring, little parental care, short life, high adult mortality, opportunistic colonizers of unstable habitats (dandelions, aphids, mice, herring). K-selected organisms maximize competitive ability near carrying capacity: large body size, late maturity, few large offspring, extensive parental care, long life, low adult mortality, equilibrial occupants of stable habitats (elephants, condors, oaks, whales).

Pianka (1970) extended the framework, and Grime (1977) added a third axis for plants — competitor / stress-tolerator / ruderal (CSR). Modern life-history theory (Stearns 1992, Roff 2002) treats these as trade-offs along Pareto fronts shaped by intrinsic constraints (energy, time, geometry) rather than discrete categories. Semelparous species reproduce once and die (Pacific salmon Oncorhynchus spp., agave, century plants, mayflies, periodical cicadas Magicicada spp. with 13-year and 17-year prime-numbered cycles); iteroparous species reproduce repeatedly (most mammals, most trees). Cole’s paradox (1954) asks why iteroparity exists at all given that an immortal semelparous strategy with one extra offspring matches an iteroparous lifetime — the resolution is mortality between breeding bouts.

2.4 Metapopulations

Levins (1969) defined a metapopulation as a “population of populations” — discrete local populations on patches of habitat connected by occasional dispersal. The Levins model tracks the fraction of patches occupied p: dp/dt = c·p(1−p) − e·p, where c is colonization rate per occupied patch and e is extinction rate per occupied patch. Equilibrium occupancy p* = 1 − e/c. The model predicts regional persistence even when every local population goes extinct eventually, provided c > e. Source–sink dynamics (Pulliam 1988) extend this to a heterogeneous landscape: source patches export migrants, sink patches have intrinsic r < 0 but are maintained by immigration. This reframes conservation: a “thriving” population in a sink habitat can be misleading evidence about the species’ true viability.

3. Demography and life-history mathematics

The Euler–Lotka equation Σ e^{−rx} lx mx = 1 implicitly defines the intrinsic rate of increase r given the life table. Leslie (1945) cast age-structured demography as matrix multiplication: n(t+1) = L·n(t), where L is the Leslie matrix with fertilities along the top row and survival probabilities along the subdiagonal. The dominant eigenvalue λ = e^r is the asymptotic growth rate, and the corresponding right eigenvector is the stable age distribution. The left eigenvector gives reproductive value v(x) — Fisher’s measure of how much an individual of age x contributes to future generations.

Big-bang reproduction (semelparity) maximizes a single bout of reproductive effort at the expense of survival. Pacific salmon are the textbook case: chinook Oncorhynchus tshawytscha, sockeye O. nerka, and four congeners migrate up natal streams, spawn, and die within days. The evolutionary explanation is that adult survival prospects are so poor (or future reproductive opportunities so reduced) that all energy is best invested in a final reproductive event — Charnov and Schaffer (1973) formalized this with the marginal-value comparison.

4. Community ecology

4.1 Interaction types

Pairwise interactions between species are classified by their effects on each partner’s fitness (per-capita growth rate):

Interaction	Effect on A	Effect on B
Competition	−	−
Predation	+	−
Parasitism	+	−
Mutualism	+	+
Commensalism	+	0
Amensalism	0	−
Neutralism	0	0

Predation and parasitism share the +/− signature but differ in time scale (one feeding event vs chronic association) and lethality (usually lethal vs usually sub-lethal). Herbivory is structurally predation on plants but is often treated as a separate category because plants tolerate partial consumption.

4.2 Lotka–Volterra equations

Predator–prey (Lotka 1925, Volterra 1926, motivated by Mediterranean fish-catch records during World War I):

dN/dt = rN − aNP dP/dt = baNP − mP

where N is prey, P is predator, r is prey growth rate, a is attack rate, b is conversion efficiency of prey into new predators, m is predator mortality. The system has a neutrally stable equilibrium at (N*, P*) = (m/(ba), r/a) and produces closed limit cycles around it — predator and prey oscillate out of phase by π/2. The Canadian lynx (Lynx canadensis) – snowshoe hare (Lepus americanus) cycle in the Hudson’s Bay Company fur-trapping records (Elton 1924) is the canonical empirical case, though modern work shows three trophic levels (plants → hares → lynx) drive it.

Competition for shared resources:

dN1/dt = r1·N1·(K1 − N1 − α12·N2)/K1 dN2/dt = r2·N2·(K2 − N2 − α21·N1)/K2

α12 is the competition coefficient — the effect of one individual of species 2 on the per-capita growth of species 1 measured in equivalent individuals of species 1. Stable coexistence requires K1/α12 > K2 AND K2/α21 > K1 (each species must inhibit itself more strongly than the other species inhibits it). Otherwise one excludes the other.

4.3 Niche, competitive exclusion, character displacement

Gause (1934) cultured Paramecium caudatum and P. aurelia separately and together; the former went extinct in mixed culture every time. This established the competitive exclusion principle: two species with identical niches cannot coexist indefinitely.

Hutchinson (1957) defined the niche as an n-dimensional hypervolume in environmental-resource space — every axis (temperature, humidity, prey size, perch height, pH, soil moisture, etc.) is a dimension, and the niche is the region of that space in which a species persists. The fundamental niche is everywhere a species could live in the absence of competitors; the realized niche is the subset actually occupied given competition, predation, and dispersal constraints. Connell’s 1961 barnacle experiment on Scottish rocky shores (Chthamalus stellatus excluded from the lower intertidal by Semibalanus balanoides but capable of growing there alone) was the classic demonstration.

Character displacement is the divergence of competing species in traits relevant to resource use where they co-occur (sympatry) compared to where they don’t (allopatry). Brown and Wilson (1956) named the pattern; the Galápagos ground finches Geospiza fortis and G. magnirostris on Daphne Major, studied by Peter and Rosemary Grant since 1973, show measurable beak-depth shifts within a single drought generation (Grant & Grant 2006, Science).

4.4 Trophic structure and energy flow

Trophic levels number upward from primary producers:

1. Primary producers — photosynthetic plants, algae, cyanobacteria; chemosynthetic bacteria at hydrothermal vents (Lonsdale 1977, Riftia pachyptila tubeworms with Endoriftia sulfide-oxidizing symbionts).
2. Primary consumers (herbivores) — grasshoppers, cattle, manatees, krill.
3. Secondary consumers (carnivores) — frogs, weasels, tuna.
4. Tertiary consumers / apex predators — eagles, sharks, killer whales, lions.
5. Decomposers / detritivores — fungi, bacteria, earthworms, vultures; technically operate at all levels by consuming dead biomass.

Lindeman (1942) measured energy flow in Cedar Bog Lake, Minnesota, and reported the 10% rule — approximately one tenth of the energy fixed at one trophic level is incorporated into biomass at the next. This sets a hard limit on food-chain length: with NPP ~10 MJ·m^−2·yr^−1 at the base, a fifth-level consumer has only ~1 kJ·m^−2·yr^−1 available, insufficient to support a sustained population of large-bodied predators. Real food webs typically have 3–5 trophic levels.

Top-down (predator-controlled) vs bottom-up (resource-controlled) regulation is a long-running debate. Hairston, Smith, and Slobodkin (1960, “HSS hypothesis”: “the world is green”) argued top-down control of herbivores by predators explains why terrestrial plants are not all eaten. Modern view: both operate, with relative strength varying by system.

Trophic cascades are indirect effects of an apex predator on lower levels via its prey. Estes and Palmisano (1974) showed sea otters Enhydra lutris in Alaska control sea urchins (Strongylocentrotus) which control kelp; otter removal collapses kelp forests. The Yellowstone wolf reintroduction (1995) is the most-cited terrestrial case: wolves Canis lupus reduce elk Cervus canadensis herbivory on willow and aspen in riparian zones, with downstream effects on beavers, songbirds, and possibly stream channel morphology (Ripple and Beschta 2012, though magnitude is debated).

4.5 Keystone species and ecosystem engineers

Paine (1969) coined “keystone species” after removing the predatory starfish Pisaster ochraceus from rocky tide pools on the Washington coast; mussel Mytilus californianus monopolized space and species richness collapsed from 15 to 8. A keystone species has impact disproportionate to its abundance.

Ecosystem engineers physically modify habitat: beavers Castor canadensis create wetlands; reef-building corals Acropora and Porites build the matrix the rest of the reef inhabits; earthworms restructure soil profiles (Darwin’s last book, 1881); prairie dogs Cynomys build burrow systems used by burrowing owls and rattlesnakes.

5. Diversity measurement

Species richness S is the count of distinct species in a sample, ignoring abundances. It is sample-size-dependent — rarefaction curves and the Chao1 estimator (Chao 1984) correct for under-sampling.

Shannon diversity H’ = −Σ pi ln pi, where pi is the proportional abundance of species i. Maximum is ln S when all species are equally abundant. Originated as an entropy measure (Shannon 1948) and was imported into ecology.

Simpson diversity D = Σ pi² is the probability that two randomly drawn individuals are the same species; 1 − D and 1/D are more commonly reported as “diversity”.

Pielou evenness J’ = H’/ln S, ranging 0–1, separates evenness from richness.

Alpha, beta, gamma diversity (Whittaker 1960): α is local within-site diversity, γ is regional total, β is the turnover between sites computed as γ/α or as a pairwise dissimilarity metric (Jaccard, Sørensen, Bray–Curtis).

Species–area relationship S = c·A^z (Arrhenius 1921, formalized by Preston 1962); z ≈ 0.25 for islands and habitat fragments, ≈0.15 for nested mainland samples. Doubling area roughly increases species count by 19% for z = 0.25.

Rank–abundance curves plot log(pi) against species rank from most to least abundant. Geometric, log-normal (Preston), broken-stick (MacArthur), and Zipf distributions correspond to different community assembly models.

6. Ecosystem ecology and biogeochemistry

6.1 Nutrient cycles

Carbon moves through reservoirs of atmosphere (CO2 + CH4), ocean (DIC ~38 000 Pg C), terrestrial vegetation and soils (~2 500 Pg), and fossil deposits (~5 000 Pg recoverable). Fluxes: photosynthesis ~120 Pg C·yr^−1 (gross primary productivity, GPP), respiration ~120 Pg, anthropogenic fossil-fuel emissions ~10 Pg·yr^−1 (2020s), cement production ~1.6 Pg. Net primary productivity (NPP) = GPP − autotrophic respiration ≈ 60 Pg·yr^−1 terrestrial, ~50 Pg·yr^−1 ocean.

Nitrogen is overwhelmingly N2 (~78% of atmosphere) but triple-bond cleavage requires either biological fixation by nitrogenase enzymes in Rhizobium, Bradyrhizobium (legume root nodules), Frankia (alder, Casuarina), and free-living cyanobacteria (Anabaena, Nostoc) — or the Haber–Bosch industrial process (Haber, Nobel 1918; Bosch, Nobel 1931) which now contributes roughly half of all reactive nitrogen entering the biosphere annually. Nitrification (Nitrosomonas NH4+ → NO2−, Nitrobacter NO2− → NO3−) and denitrification (Pseudomonas, Paracoccus NO3− → N2) close the cycle. Anthropogenic excess causes eutrophication, hypoxia (Gulf of Mexico dead zone ~15 000 km²), and nitrous oxide release (N2O is a 273× CO2-equivalent greenhouse gas over a 100-year horizon, IPCC AR6).

Phosphorus has no atmospheric gas phase (PH3 is trace); cycle is dominated by weathering of phosphate rock (apatite Ca5(PO4)3(F,Cl,OH)) and biological uptake/release. Limiting nutrient in most freshwater systems and in old tropical soils. Eutrophication of Lake Erie in the 1960s was driven by detergent phosphate.

Sulfur cycles via volcanic SO2, sulfate aerosols, DMS from marine algae (Charlson–Lovelock–Andreae–Warren CLAW hypothesis 1987 linking marine DMS to cloud condensation nuclei), bacterial sulfate reduction and sulfide oxidation.

Water cycle: evaporation, transpiration (combined ET ~70 × 10³ km³·yr^−1 globally), precipitation, runoff, groundwater flow. Residence times: atmosphere ~9 days, rivers ~2 weeks, soil moisture ~2 months, lakes ~years to centuries, deep groundwater ~10⁴ years, oceans ~3 200 years.

See biochemistry-foundations for enzymatic detail on nitrogen fixation and photosynthesis, and environmental-engineering for engineered nutrient removal.

6.2 Productivity, eutrophication, regime shifts

Primary productivity is measured by ¹⁴C uptake (Steemann Nielsen 1952), oxygen evolution in light/dark bottles, or remote-sensing chlorophyll-a (SeaWiFS, MODIS, since 1997). NPP varies from <0.1 mol C·m^−2·yr^−1 (open ocean gyres) to >100 (tropical rainforest, salt marsh, coral reef).

Eutrophication driven by N + P loading shifts shallow lakes between clear-water macrophyte-dominated and turbid algae-dominated alternative stable states (Scheffer et al. 1993). Hysteresis means reducing nutrients to the level that initially caused the shift will not reverse it; deeper cuts are required. This is the prototype “regime shift” in ecology.

7. Biogeography and biomes

Wallace (1876) mapped the major biogeographic realms — Nearctic, Neotropical, Palearctic, Afrotropical, Oriental, Australasian, Oceanic, Antarctic — separated by deep-history barriers including the Wallace Line through the Indonesian archipelago that marks the boundary between Asian and Australasian faunas.

Latitudinal diversity gradient — species richness peaks in the tropics and declines toward the poles for almost every taxon. Hypotheses include greater time-integrated stability and higher productivity in the tropics, niche-conservatism (most clades originated tropical and few have invaded cold climates), faster evolutionary rates in the tropics (Mittelbach et al. 2007). No consensus.

Island biogeography (MacArthur and Wilson 1967): equilibrium species number on an island balances immigration (decreasing with distance from source) against extinction (decreasing with area). S* depends on area and isolation. Empirical z values for the species-area relationship match predictions. The framework was extended to “habitat islands” — forest fragments in agricultural matrices, mountaintop sky islands — and underpins reserve-design principles (SLOSS: single large or several small).

Köppen–Geiger climate classification (Köppen 1884, Geiger updates 1954, 1961) maps climates by temperature and precipitation regime into A (tropical), B (arid), C (temperate), D (continental), E (polar), with finer subdivisions. Biome boundaries align well with Köppen classes — Holdridge life zones (1947) refine this with biotemperature and PET.

Major biomes:

• Tropical rainforest — equatorial, >2000 mm rain·yr^−1, mean T > 20 °C; highest biodiversity per hectare (Amazon, Congo, Indo-Malayan); NPP 20–35 Mg·ha^−1·yr^−1.
• Savanna — seasonally dry tropics, scattered trees + grass; fire-maintained; Serengeti, Cerrado, Llanos.
• Desert — <250 mm rain·yr^−1; Sahara, Atacama (driest non-polar), Gobi, Sonoran.
• Mediterranean / chaparral — winter-rainy, summer-dry; California, Chile, Cape, southwest Australia, Mediterranean basin; high endemism.
• Temperate deciduous forest — eastern North America, Europe, East Asia; 4 seasons; NPP ~12 Mg·ha^−1·yr^−1.
• Temperate grassland — North American prairie, Eurasian steppe, pampas; mostly converted to agriculture (corn belt, wheat belt).
• Boreal forest / taiga — Canada, Scandinavia, Siberia; conifer-dominated (Picea, Pinus, Abies, Larix); largest contiguous biome.
• Tundra — permafrost, low NPP, short growing season; arctic and alpine.
• Marine pelagic — open ocean, photic zone 0–200 m drives primary production by phytoplankton (diatoms, dinoflagellates, coccolithophores like Emiliania huxleyi).
• Marine benthic — sea floor; abyssal plains, mid-ocean ridges, hydrothermal vents.
• Intertidal — between high and low tide; zonation studied by Connell, Paine; gradient of physical stress (drying) and biological stress (predation).
• Coral reef — shallow tropical seas; Symbiodinium dinoflagellate endosymbionts in scleractinian coral hosts; highest marine biodiversity.

8. Conservation biology

8.1 Extinction rates and biodiversity loss

Background extinction rate inferred from the fossil record is roughly 1 extinction per million species-years (Pimm et al. 1995, refined Pimm 2014 Science) — so a typical species lifespan is ~1 Myr. Current observed extinction rate among well-monitored taxa (birds, mammals, amphibians) is 100–1000× the background, qualifying as a sixth mass extinction event (Ceballos, Ehrlich, Barnosky 2015).

IUCN Red List categories: Extinct (EX), Extinct in the Wild (EW), Critically Endangered (CR), Endangered (EN), Vulnerable (VU), Near Threatened (NT), Least Concern (LC), Data Deficient (DD), Not Evaluated (NE). CR/EN/VU together are “threatened”. Quantitative criteria use population decline rate, geographic range, population size, and probability of extinction.

8.2 Threats and frameworks

HIPPO (E. O. Wilson) — the five dominant drivers of biodiversity loss, by approximate importance:

• Habitat loss / fragmentation — deforestation, conversion to agriculture, urbanization.
• Invasive species — Boiga irregularis brown tree snake on Guam, Salvinia molesta in tropical wetlands, Rhinella marina cane toad in Australia, zebra mussel Dreissena polymorpha, chytrid fungus Batrachochytrium dendrobatidis in amphibians.
• Pollution — chemical, nutrient, light, noise, microplastic.
• Population (human) — driver of all the others.
• Overharvesting — fisheries (cod Gadus morhua collapse 1992 Newfoundland), bushmeat trade, pet trade, ornamental plants.

Biodiversity hotspots (Myers et al. 2000, Nature; updated 2004 to 36 hotspots): regions with >1500 endemic vascular plant species and >70% original habitat loss. Together they cover ~2.4% of Earth’s land surface but contain >50% of endemic plants and ~43% of endemic terrestrial vertebrates. Conservation funding prioritization tool.

8.3 De-extinction and rewilding

De-extinction via cloning (Pyrenean ibex Capra pyrenaica pyrenaica, briefly resurrected 2003), back-breeding (aurochs from cattle), or genome editing of close relatives (woolly mammoth-like edits in Asian elephant by Colossal Biosciences) is technologically partial and ecologically and ethically contested.

Rewilding restores ecosystems via reintroduction of extirpated keystone species and removal of management pressure — Oostvaardersplassen (Netherlands), Yellowstone wolves, European bison Bison bonasus in the Carpathians, beaver reintroductions in Britain since 2009.

8.4 Climate change ecology

Phenological mismatch: oak Quercus robur bud-burst, winter-moth Operophtera brumata caterpillar emergence, and great tit Parus major egg-laying in the Netherlands have de-synchronized as spring warms unevenly (Visser et al. 2006). Range shifts: poleward at ~17 km·decade^−1 and upslope ~11 m·decade^−1 (Chen et al. 2011, Science). Coral bleaching: when sea-surface temperature exceeds the local long-term summer mean by ~1 °C for several weeks, corals expel Symbiodinium and bleach; Great Barrier Reef mass bleaching events 1998, 2002, 2016, 2017, 2020, 2022. Ocean acidification: surface pH has dropped from ~8.21 (preindustrial) to ~8.10 today, with calcium carbonate saturation state declining and threatening calcifiers (corals, coccolithophores, pteropods, oysters). IPCC AR6 (2021, 2022) synthesizes the projections.

9. Evolutionary biology — core theory

9.1 Darwin and Wallace

Charles Darwin and Alfred Russel Wallace presented joint papers at the Linnean Society on 1 July 1858. Darwin’s On the Origin of Species by Means of Natural Selection (1859) presented the argument as: (1) populations produce more offspring than can survive; (2) individuals vary in heritable traits; (3) variants better suited to the local environment leave more offspring; (4) over generations, populations change. The mechanism is undirected — variation arises blind to need, selection sieves it.

Darwin’s The Descent of Man, and Selection in Relation to Sex (1871) introduced sexual selection — selection arising from differential mating success rather than differential survival — explaining traits that reduce viability (peacock Pavo cristatus tail, bird of paradise plumage, elk antlers).

9.2 The Modern Synthesis (1930s–1940s)

Population genetics reconciled Mendelian inheritance with continuous variation through the work of R. A. Fisher (The Genetical Theory of Natural Selection, 1930), J. B. S. Haldane (The Causes of Evolution, 1932), and Sewall Wright (1931, 1932 on shifting balance and the adaptive landscape). Dobzhansky’s Genetics and the Origin of Species (1937), Mayr’s Systematics and the Origin of Species (1942, biological species concept), Simpson’s Tempo and Mode in Evolution (1944, integrating paleontology), and Stebbins’ Variation and Evolution in Plants (1950) extended the synthesis across organisms.

9.3 Hardy–Weinberg equilibrium

For a single autosomal locus with alleles A (frequency p) and a (frequency q = 1−p): allele frequencies remain constant and genotype frequencies are p² (AA) : 2pq (Aa) : q² (aa) under the assumptions of (1) random mating, (2) no selection, (3) no mutation, (4) no migration, (5) infinite population size (no drift). Hardy and Weinberg independently derived this in 1908. Deviations from H–W expectations diagnose evolution.

9.4 Forces of evolutionary change

• Mutation — ultimate source of variation. Human germline mutation rate ~1.2 × 10⁻⁸ per base per generation (Kong et al. 2012, Nature), giving ~70 new mutations per zygote across 3 × 10⁹ bp. Mutation rate alone changes allele frequencies very slowly.
• Selection — differential reproduction by genotype. Selection coefficient s = 1 − w, where w is relative fitness. Change in allele frequency per generation Δq ≈ −s·p·q²/(1 − s·q²) for a recessive deleterious allele. Strong selection (s > 0.1) can fix or eliminate alleles in tens of generations; weak selection (s ~ 10⁻⁴) operates on geological time scales.
• Genetic drift — random sampling of alleles each generation. Variance in allele frequency per generation ≈ pq/(2Ne) where Ne is the effective population size. Heterozygosity decays as Ht = H0 (1 − 1/(2Ne))^t ≈ H0 e^{−t/(2Ne)}. Probability of fixation of a new neutral mutation = 1/(2Ne) and expected time to fixation conditional on fixation = 4Ne generations.
• Gene flow / migration — movement of alleles between populations at rate m homogenizes them; FST ≈ 1/(1 + 4Nem) for an island model (Wright). One migrant per generation (Nem = 1) is roughly sufficient to prevent strong differentiation.
• Non-random mating — assortative mating, inbreeding, sexual selection. Inbreeding coefficient F measures probability of identity by descent; F > 0 reduces heterozygosity by a factor (1 − F).

9.5 Types of selection

By direction:

• Directional — shifts trait mean (Darwin’s finches in drought, antibiotic resistance evolution).
• Stabilizing — preserves trait mean, reduces variance (human birth weight, optimal egg number).
• Disruptive / diversifying — favours extremes, can split a population (jaw morphology in cichlid trophic specialists).

By sign:

• Positive selection — increase in beneficial allele frequency; selective sweep produces hitchhiking and reduced diversity at linked sites.
• Negative / purifying selection — removes deleterious mutations, the dominant mode at conserved sites.
• Balancing selection — maintains polymorphism: overdominance (heterozygote advantage — sickle-cell allele HbS gives malaria resistance in heterozygotes, lethality in homozygotes — Allison 1954), negative frequency-dependent (rare-morph advantage in scale-eating cichlid Perissodus microlepis left/right mouth asymmetry; MHC alleles).

Sexual selection mechanisms include intrasexual combat (male–male competition for mates: elephant seal harems, red deer roars and antler fights) and intersexual choice (female mate choice for ornaments). Fisherian runaway (Fisher 1930) shows preference and ornament co-evolve as long as they remain genetically correlated. Good-genes models (Zahavi handicap principle 1975, Iwasa–Pomiankowski Heredity 1991) require the ornament to honestly signal mate quality at a cost.

9.6 Neutral and nearly neutral theory

Motoo Kimura (1968 Nature, 1983 book) proposed that most molecular evolutionary change is neutral or near-neutral and driven by drift, not selection. Tomoko Ohta (1973) extended to nearly neutral theory in which slightly deleterious mutations are effectively neutral when |Nes| < 1. This is the null model against which positive selection is tested: McDonald–Kreitman test (1991), dN/dS ratio at codon sites (Yang and Nielsen), Tajima’s D (1989), HKA test, extended haplotype homozygosity (Sabeti et al. 2002).

9.7 Coalescent and molecular clock

Kingman (1982) introduced the coalescent — looking backward in time, lineages in a sample merge with rate (k choose 2)/(2Ne) when k lineages exist. Expected time to coalescence for a pair is 2Ne generations; expected time to MRCA of n samples is 4Ne(1 − 1/n).

Molecular clock — Zuckerkandl and Pauling 1962 noted approximately constant amino-acid substitution rates among lineages, allowing divergence-time estimation from sequence divergence given a calibration. Rate heterogeneity across lineages and sites required relaxed-clock models in modern Bayesian phylogenetics (BEAST 2, MrBayes 3).

10. Speciation

10.1 Modes

• Allopatric — population split by geographic barrier (mountain rise, river capture, glacial vicariance), each evolves independently, accumulating reproductive isolation. The default mode of animal speciation under the Mayr framework. Galápagos finch radiation (Darwin’s finches, 18 species derived from a continental ancestor 1.5–2 Mya), Hawaiian honeycreepers, African cichlids in Lake Tanganyika and Lake Victoria.
• Peripatric — small peripheral isolate undergoes rapid divergence (founder effect speciation, Mayr 1954). Less mainstream now than in mid-century but invoked for island endemics.
• Parapatric — populations along a continuous gradient diverge despite ongoing limited gene flow; selection along an environmental cline overcomes migration. Mine-tolerance ecotypes in grasses (Agrostis tenuis on copper-mine tailings, Antonovics).
• Sympatric — divergence within a single geographically continuous population. Long contested; clearest cases involve host shifts in phytophagous insects (apple maggot fly Rhagoletis pomonella shifted from native hawthorn Crataegus to introduced apples in ~1860, host-associated divergence Feder and Bush), and trophic polymorphisms in fish in single crater lakes (cichlids in Lake Apoyo, Nicaragua, Barluenga et al. 2006).

10.2 Reproductive isolation

Prezygotic barriers: ecological isolation, temporal isolation (mating-time differences), behavioral isolation (courtship signals), mechanical isolation (genitalia mismatch), gametic isolation. Postzygotic barriers: hybrid inviability, hybrid sterility (mule, hinny — Equus caballus × E. asinus hybrids), hybrid breakdown in F2 / backcross.

Dobzhansky–Muller incompatibilities — alleles that function in their own genetic background but interact deleteriously in hybrids. Each lineage fixes new alleles that are individually neutral or beneficial at home but combine into broken epistatic networks abroad. Explains why hybrid sterility evolves without any “valley” being crossed.

Hybrid zones — narrow regions where two diverged populations meet and produce hybrids. Cline width depends on selection against hybrids and dispersal. Bombina bombina × B. variegata toad hybrid zone in Europe is a textbook case.

Ring species — a chain of populations connected by gene flow around a barrier with terminal populations that, where they meet, behave as reproductively isolated species. Ensatina eschscholtzii salamanders around California’s Central Valley (Stebbins 1949, Wake et al.) and greenish warbler Phylloscopus trochiloides around Tibetan Plateau (Irwin et al. 2001) are the most-cited examples, though both have come under scrutiny.

11. Phylogenetics and tree of life

Willi Hennig’s Grundzüge einer Theorie der phylogenetischen Systematik (1950, English translation 1966) established cladistics: classification by shared derived characters (synapomorphies), not overall similarity. Monophyletic groups (a clade — common ancestor + all descendants) are the only valid taxa. Methods to infer trees:

• Parsimony — minimize total character-state changes.
• Maximum likelihood — find the tree that maximizes P(data | tree, model); standard tools include RAxML, IQ-TREE.
• Bayesian inference — posterior over trees via MCMC; BEAST 2, MrBayes 3, RevBayes.
• Distance methods — Neighbor joining (Saitou and Nei 1987), UPGMA.

Substitution models (JC69, K80, HKY85, GTR, with rate variation Γ and proportion invariant I) account for unequal substitution rates and the underestimation of divergence at multiple-hit sites.

The three-domain tree (Carl Woese 1977, 1990, comparative 16S/18S rRNA sequencing) split prokaryotes into Bacteria and Archaea, with Eukarya a sister or nested clade. Modern eocyte / two-domain hypotheses (Embley, Williams 2015) place eukaryotes within Archaea, sister to Asgard archaea (Lokiarchaeota), strengthening the case for an archaeal host in eukaryogenesis.

LUCA (Last Universal Common Ancestor) is reconstructed as a chemolithotrophic, anaerobic, thermophilic organism with a near-modern genetic code, ~3.9–4.1 Gya (Weiss et al. 2016 Nat Microbiol).

Eukaryogenesis — Lynn Margulis (then Sagan) 1967 J Theor Biol revived and formalized endosymbiotic theory: mitochondria descended from an α-proteobacterium (closest extant relatives in Rickettsiales), plastids from a cyanobacterium captured by a heterotrophic eukaryote (Cavalier-Smith and others). Secondary endosymbiosis explains complex plastids in stramenopiles, dinoflagellates, euglenids, apicomplexans. Horizontal gene transfer (HGT) is rampant in prokaryotes (transformation, transduction, conjugation) and complicates “tree” thinking — Doolittle’s “web of life”.

12. Major evolutionary transitions

Maynard Smith and Szathmáry’s The Major Transitions in Evolution (1995) identified eight transformations in which entities that previously replicated independently became combined into a higher-level replicator:

1. Replicating molecules → populations of molecules in compartments
2. Independent replicators → chromosomes
3. RNA as both gene and enzyme → DNA + protein with the genetic code
4. Prokaryotes → eukaryotes
5. Asexual clones → sexual populations
6. Unicellular protists → multicellular organisms (independently in at least 25 lineages)
7. Solitary individuals → colonies with sterile castes (eusociality in Hymenoptera, termites, naked mole-rats Heterocephalus glaber)
8. Primate societies → human societies with language

Each transition required mechanisms to prevent within-group conflict (replicator suppression, kin recognition, policing).

13. Evo-devo

Evolutionary developmental biology (evo-devo) emerged in the 1980s with the discovery that Hox genes — a cluster of homeobox-containing transcription factors specifying segmental identity along the anterior–posterior axis — were conserved between Drosophila, mouse, and beyond (E. B. Lewis, Wieschaus, Nüsslein-Volhard, Nobel 1995). Anterior-to-posterior gene order in the cluster corresponds to anterior-to-posterior expression in the embryo (colinearity).

Body plans (Bauplans) of the major animal phyla appeared near-simultaneously in the Cambrian explosion ~541–518 Mya, documented in the Burgess Shale (Charles Walcott 1909, British Columbia) and the older Chengjiang Lagerstätte in Yunnan (1984). The pre-Cambrian Ediacaran biota (~575–541 Mya) shows soft-bodied forms whose phylogenetic affinities are still debated. Deep homology — the use of orthologous regulatory genes for analogous structures in distantly related lineages (Pax6 in all eyes, Distal-less in all appendages) — was named by Shubin, Tabin, and Carroll (1997). Modularity of developmental gene regulatory networks (GRNs, Davidson) allows tinkering: cooption of an old module for a new structure, redeployment in a new tissue.

14. Macroevolution and mass extinctions

Stephen Jay Gould and Niles Eldredge (1972) proposed punctuated equilibria — long stasis interrupted by rapid morphological change at speciation events — as an alternative to gradualism. Tempo controversies persist; data support a mixture of patterns by lineage and trait.

Mass extinctions are intervals with extinction rates far above background. Raup and Sepkoski (1982) identified the Big Five from marine invertebrate genus turnover:

1. End-Ordovician ~444 Mya — ~85% species lost; glaciation as Gondwana drifted over the South Pole, sea level fall, anoxia.
2. Late Devonian ~372–359 Mya — multiple pulses (Kellwasser and Hangenberg events); ~75% species lost; anoxia, possibly plant-driven nutrient pulses.
3. End-Permian ~252 Mya — ~96% of marine species, ~70% of terrestrial vertebrate genera; the largest in Earth history; Siberian Traps flood-basalt volcanism, CO2 release, ocean acidification, anoxia, methane release.
4. End-Triassic ~201 Mya — ~80% species lost; Central Atlantic Magmatic Province (CAMP) flood basalts; cleared the stage for dinosaur dominance.
5. End-Cretaceous (K–Pg) ~66 Mya — ~76% species lost; Chicxulub impactor (~10 km diameter) struck the Yucatán; iridium anomaly at the K–Pg boundary (Luis and Walter Alvarez, Asaro, Michel 1980 Science), shocked quartz, soot, tsunami deposits; non-avian dinosaurs, ammonites, ~half of marine plankton genera went extinct.

The current sixth (Anthropocene) extinction has anthropogenic drivers (HIPPO) at rates 100–1000× background and is ongoing.

15. Human evolution

Hominin fossils document a ~7-Myr lineage from the chimpanzee–human last common ancestor:

• Sahelanthropus tchadensis ~7 Mya, Chad, possibly bipedal cranium.
• Orrorin tugenensis ~6 Mya, Kenya.
• Ardipithecus ramidus ~4.4 Mya, Ethiopia.
• Australopithecus afarensis ~3.9–2.9 Mya — “Lucy” AL 288-1 discovered by Donald Johanson 1974, Hadar, Ethiopia; bipedal locomotion confirmed by Laetoli footprints (Mary Leakey 1978, 3.66 Mya).
• Australopithecus africanus ~3.3–2.1 Mya, southern Africa; Taung child (Dart 1924).
• Paranthropus robustines — robust megadont side branch.
• Homo habilis ~2.4–1.4 Mya, larger brain, Oldowan stone tools.
• Homo erectus ~2 Mya–110 kya, first to leave Africa, Acheulean hand-axes, possibly first controlled fire.
• Homo heidelbergensis ~700–200 kya, presumed common ancestor of Neanderthals, Denisovans, and modern humans.
• Homo neanderthalensis ~400–40 kya, Eurasia.
• Homo sapiens ~300 kya — Jebel Irhoud, Morocco (Hublin et al. 2017 Nature), the earliest currently accepted modern human fossils; Omo Kibish (Ethiopia) 195 kya.

Svante Pääbo and the Max Planck Institute for Evolutionary Anthropology recovered the Neanderthal nuclear genome (2010), the Denisovan genome from a Siberian phalanx (Reich et al. 2010), and showed non-African modern humans carry 1–4% Neanderthal ancestry and Melanesians carry up to ~5% Denisovan ancestry from introgression. Pääbo received the Nobel Prize in Physiology or Medicine 2022 for paleogenomics.

16. Behavioral ecology

16.1 Optimal foraging

MacArthur and Pianka (1966) framed foraging as an economic optimization: include a prey type if its energy/handling-time ratio exceeds the average for the foraging environment if it is excluded. Charnov’s marginal value theorem (1976) predicts patch-leaving times when patches deplete: leave when instantaneous intake rate falls to the long-term average across the habitat. Empirical tests with bumblebees, starlings, and great tits broadly confirm.

16.2 Game theory and ESS

John Maynard Smith and George Price (1973, 1974) imported evolutionary game theory. An evolutionarily stable strategy (ESS) is a strategy that, if adopted by the population, cannot be invaded by any alternative. The hawk–dove game has hawks always fight (cost C if both hawks meet), doves always display and yield — mixed ESS frequency of hawks = V/C where V is contested resource value. Producer–scrounger, ideal-free distribution, war of attrition, and parent–offspring conflict are other canonical games.

16.3 Kin selection and altruism

W. D. Hamilton (1964 J Theor Biol) formalized inclusive fitness: an altruistic act spreads if rB > C, where r is the coefficient of relatedness, B is fitness benefit to the recipient, and C is fitness cost to the actor. Eusociality in haplodiploid Hymenoptera (workers share 3/4 of genes with sisters, only 1/2 with own offspring) was a famous early prediction, though haplodiploidy is neither necessary (termites are diploid) nor sufficient (many haplodiploids are not eusocial).

Robert Trivers (1971 Q Rev Biol) introduced reciprocal altruism — repeated interactions allow cooperation among non-kin when partners can track histories and defect against defectors. Trivers (1972) wrote the foundational paper on parental investment, predicting that the sex investing more in offspring becomes the choosier and limiting sex. Trivers (1974) wrote on parent–offspring conflict — offspring favoured greater investment than parents are selected to provide.

Hamilton–Zuk hypothesis (1982): female preference for elaborate male traits arises because such traits honestly signal parasite resistance. Empirical support is mixed.

16.4 Group selection debates

V. C. Wynne-Edwards (1962) argued populations regulate themselves through self-limited breeding for the good of the group. G. C. Williams (Adaptation and Natural Selection, 1966) demolished the argument: individual-level selection swamps group-level selection in almost all realistic parameter ranges. Multi-level selection (D. S. Wilson and E. Sober Unto Others, 1998) revived a quantitatively careful form using the Price equation to partition selection into within- and between-group components. The mathematical equivalence between multi-level selection and inclusive fitness (Marshall, Queller) ended much of the heat; the disagreement is now largely about which framework yields cleaner empirical work.

E. O. Wilson’s Sociobiology: The New Synthesis (1975) extended evolutionary explanation to vertebrate (including human) behaviour and triggered the sociobiology controversy. Evolutionary psychology (Cosmides, Tooby, Buss from late 1980s) continues that programme for human cognition.

17. Field and quantitative methods

• Mark–recapture — Lincoln–Petersen, Schnabel, Jolly–Seber, robust design (Pollock 1982), Cormack–Jolly–Seber for survival.
• Quadrats and transects — standard sampling for sessile or slow-moving organisms.
• Camera traps — non-invasive population estimation, occupancy modelling (MacKenzie et al. 2002).
• Acoustic monitoring — passive autonomous recording units for cetaceans (sperm whale Physeter macrocephalus click trains), bats, songbirds, anurans. Machine-learning classifiers (BirdNET, Cornell Lab) now process millions of hours.
• eDNA metabarcoding — environmental DNA shed into water or soil is amplified at standard markers (12S, COI, 16S, ITS) and sequenced on Illumina MiSeq / NextSeq or Oxford Nanopore platforms (PacBio Sequel II / Revio for full-length amplicons). Detects rare and cryptic species.
• Satellite remote sensing — Landsat (since 1972), MODIS, Sentinel-2, GEDI lidar (since 2018) for canopy structure, NDVI for productivity, SeaWiFS / MODIS-Aqua for ocean chlorophyll.

Modelling frameworks include compartmental ODEs (predator–prey, SIR), individual-based / agent-based models (NetLogo, Mesa), network models (food-web stability after May 1972; degree distribution analysis), Bayesian hierarchical state-space models (Stan, JAGS, NIMBLE), occupancy models (PRESENCE, unmarked R package), demographic projection matrices (Caswell 2001). See probability-fundamentals, bayesian-inference, ode-numerical-methods, graph-theory.

18. Quantitative genetics

Beyond the single-locus framework, quantitative genetics models continuous traits as the sum of many small-effect loci plus environment. The breeder’s equation R = h² · S relates response to selection R, narrow-sense heritability h² (= additive genetic variance / total phenotypic variance), and the selection differential S (the difference between selected parents and the unselected population mean). Falconer and Mackay’s Introduction to Quantitative Genetics (4th ed. 1996) is the canonical text. The animal model in mixed-effects form (BLUP — best linear unbiased prediction; Henderson 1973) underpins modern livestock breeding and is now standard in wild-population evolutionary studies through pedigreed long-term studies of Soay sheep (St Kilda), red deer (Rum), great tits (Wytham Wood, Hoge Veluwe), collared flycatchers (Gotland), and Darwin’s finches (Daphne Major).

Heritability is a population-and-environment specific summary, not a fixed property of a trait. h² for human height in well-fed Western populations is ~0.8; h² for the same trait collapses in nutrition-limited populations because environmental variance dominates. Genome-wide association studies (GWAS) on millions of human genomes have identified thousands of SNPs explaining only a fraction of estimated heritability (“missing heritability” — Manolio et al. 2009 Nature); rare variants, epistasis, gene–environment interactions, and overestimation of pedigree-based h² each contribute.

QTL mapping identifies regions of the genome that contribute to trait variation in experimental crosses; GWAS does the same in unrelated population samples using linkage disequilibrium. Genomic prediction sums SNP effects into a polygenic score — used in plant and animal breeding for decades, and now in human disease risk stratification.

19. Spatial ecology and landscape

The landscape is a mosaic of patches whose configuration affects population persistence and gene flow. Metrics: patch size distribution, edge density, connectivity, fractal dimension, contagion. Tools: FRAGSTATS, R packages landscapemetrics, sf, terra, raster, GIS overlays in QGIS or ArcGIS.

Corridors vs stepping stones vs matrix permeability define connectivity strategies. The Y2Y (Yellowstone-to-Yukon) and African transboundary parks are continental-scale connectivity initiatives.

Spatially explicit population models (RAMAS, HexSim) combine demography with movement on a habitat grid. Resistance surfaces and circuit theory (McRae 2006) compute least-cost or current-flow paths through heterogeneous landscapes for landscape genetics.

Species distribution models (SDMs) — MaxEnt, GLM/GAM, random forest, boosted regression trees — project current and future ranges from occurrence data and environmental rasters (WorldClim, CHELSA, MERRAClim). Future projections under RCP/SSP climate scenarios drive much of conservation planning under climate change.

20. Disease ecology and epidemiology of wildlife

Diseases shape ecological communities. White-nose syndrome (Pseudogymnoascus destructans, fungal) has killed tens of millions of bats in North America since 2006. Chytridiomycosis (Batrachochytrium dendrobatidis) has driven dozens of amphibian species to extinction. Devil facial tumour disease (DFTD), a transmissible cancer, has reduced Tasmanian devil Sarcophilus harrisii populations by >80%. Sea-star wasting disease devastated Pisaster ochraceus on the Pacific coast 2013–2015.

R0 (basic reproduction number) — the expected number of secondary cases from one infectious case in a fully susceptible population — separates outbreaks (R0 < 1, fades; R0 > 1, grows). Vaccination threshold for herd immunity = 1 − 1/R0 (e.g. measles R0 ~15, threshold ~93%). SIR/SEIR compartmental models (Kermack–McKendrick 1927) underpin both wildlife and human disease ecology. See ode-numerical-methods for numerical integration of compartmental ODEs.

21. Synthesis — the eco-evo loop

Ecology and evolution operate on overlapping timescales. Rapid evolution within ecological time (Hairston et al. 2005, Yoshida et al. 2003 with rotifer–algae chemostats; Reznick’s Trinidad guppy translocations 1976–present) feeds back into community dynamics: prey evolution alters predator population dynamics; predator behavioral response alters selection on prey; herbivore evolution alters plant community composition. Eco-evolutionary dynamics (Pelletier, Garant, Hendry 2009; Schoener 2011) is now a major research front. The historical separation of “fast” ecology and “slow” evolution does not survive modern data.

A few empirical landmarks make the loop concrete:

• Soapberry bugs Jadera haematoloma evolved shorter beaks within decades to feed on introduced flat-podded balloon vines (Carroll and Boyd 1992).
• Antibiotic resistance evolves on timescales of months to years and reshapes hospital microbial communities (see microbiology-foundations).
• Industrial melanism in peppered moths Biston betularia (Kettlewell 1955) rose from <1% to >95% dark morph in industrial Britain 1850–1900, and reversed after the Clean Air Acts.
• HIV-1 evolves within a single host over months under antiretroviral selection; viral phylogenies on the scale of weeks now guide outbreak forensics.
• Daphnia resurrection ecology — hatching diapausing eggs from dated sediment cores — directly measures evolution across decades.
• COVID-19 variant emergence (Alpha, Delta, Omicron lineages 2020–2022) was a real-time demonstration of eco-evolutionary dynamics in a host–pathogen system with massive surveillance data.

The unification has practical consequences. Fisheries management must account for evolution of earlier maturation and smaller body size under size-selective harvest (Olsen et al. 2004, Nature, on Atlantic cod). Pesticide and herbicide resistance evolves predictably wherever single-mode-of-action chemicals are deployed at scale. Designing vaccines and antibiotics requires anticipating pathogen evolution. Conservation under climate change requires distinguishing rescue by plasticity, dispersal, and adaptation. The eco-evolutionary view replaces the optimistic “balance of nature” with a dynamic equilibrium constantly re-set by selection acting on standing and new variation.

Adjacent

• cell-molecular-biology — the cellular substrate on which selection acts
• genetics-and-genomics — variation, inheritance, allele frequency change
• microbiology-foundations — microbial ecology, horizontal gene transfer, microbial life-history
• immunology-foundations — host–parasite coevolution at the cellular level
• biochemistry-foundations — biogeochemical cycle enzymology
• environmental-engineering — engineered nutrient and water management
• probability-fundamentals — stochastic models for populations
• graph-theory — food web topology and metapopulation networks