Module 2: Animal-Plant Chemical Interactions

2.1 Pollinator Attractants: The Chemistry of Attraction

Pollination is a mutualism mediated by chemistry: plants offer nectar (sugar solution) and pollen (protein-rich) as rewards, while pollinators provide the transfer service. The specific chemistry of nectar — its sugar composition, amino acid content, and secondary metabolite additives — is finely tuned to the sensory preferences and biophysical constraints of each pollinator guild.

Nectar Sugar Composition

Floral nectar is primarily composed of three sugars — sucrose, glucose, and fructose — in ratios that vary systematically by pollinator type:

• Hummingbird-pollinated flowers: sucrose-dominant (sucrose:hexose ratio \(\approx 4:1\)), concentration 20-25% w/w. Hummingbirds have sucrase in their gut and prefer sucrose for its higher energy per osmole.
• Bee-pollinated flowers: balanced or hexose-rich (sucrose:hexose \(\approx 1:1\)), concentration 30-50% w/w. Bees have invertase in saliva that cleaves sucrose; higher concentration maximizes energy per foraging trip.
• Butterfly-pollinated flowers: sucrose-rich (sucrose:hexose \(\approx 2:1\)), concentration 15-25% w/w. Longer proboscides penalize viscosity more.
• Bat-pollinated flowers: hexose-dominant (sucrose:hexose \(\approx 0.3:1\)), concentration 15-20% w/w. Bats drink rapidly and metabolize hexoses directly.

UV Nectar Guides

Many bee-pollinated flowers display ultraviolet patterns invisible to the human eye but strikingly visible to bees, whose photoreceptors detect UV (344 nm peak), blue (436 nm), and green (556 nm). These UV "nectar guides" are created by differential deposition of flavonoids (particularly flavonols and flavones) in petal epidermal cells. Flavonoids absorb strongly at 300-380 nm due to their conjugated ring system:

The "bullseye" pattern — UV-absorbing center (dark to bees) surrounded by UV-reflecting petals (bright to bees) — guides pollinators directly to the nectar reward, reducing handling time and increasing pollination efficiency.

Optimal Nectar Concentration from Poiseuille Flow

The rate at which a pollinator can extract nectar through its proboscis is governed by the Hagen-Poiseuille equation for laminar flow through a tube:

where \(Q\) is the volumetric flow rate (mL/s), \(r\) is the proboscis radius, \(\Delta P\) is the suction pressure, \(\eta\) is the dynamic viscosity of nectar, and \(L\) is the proboscis length.

The energy intake rate is the product of volumetric flow and energy density:

where \(\rho_E(c)\) is the energy density (kJ/mL) and \(\eta(c)\) is the viscosity, both functions of sugar concentration \(c\). Energy density increases linearly with concentration, but viscosity increases exponentially:

where \(\alpha \approx 0.17\) kJ/(mL %) and \(\beta \approx 0.06\) /%. The energy intake rate becomes:

Differentiating and setting \(d\dot{E}/dc = 0\):

For \(\beta = 0.06\)/%, the predicted optimal concentration is\(c^* \approx 17\%\) w/w for a standard proboscis. In practice, the optimal concentration depends on proboscis length: longer proboscides (moths, hummingbirds) have higher viscosity penalties (because \(Q \propto 1/L\)), shifting the optimum to lower concentrations (~15-20%), while shorter proboscides (bees) tolerate higher concentrations (~30-50%). This biophysical prediction matches observed nectar concentrations remarkably well (Kingsolver & Daniel 1983).

For more on the biophysics of proboscis feeding and flight energetics in pollinators, see our Bee Biophysics: Flight Aerodynamics module.

2.2 Herbivore Deterrents: The Chemistry of Defense

Plants deploy three major classes of chemical defenses against herbivores: tannins (quantitative defenses that reduce digestibility), alkaloids (qualitative defenses that target specific neural/metabolic targets), and cyanogenic glycosides (triggered defenses that release toxic HCN on tissue damage).

Tannins: Protein Precipitants

Tannins are high-molecular-weight polyphenols (500-20,000 Da) that bind and precipitate dietary proteins, reducing their digestibility by 30-80%. The two main classes are:

• Condensed tannins (proanthocyanidins): polymers of flavan-3-ol units (catechin, epicatechin) linked by C4-C8 or C4-C6 bonds. Found in oak bark, tea, and legume seeds.
• Hydrolyzable tannins: esters of gallic acid or ellagic acid with a glucose core. Hydrolyzed by gut tannase enzymes in some specialist herbivores.

The tannin-protein binding follows a cooperative binding model. For \(n\) independent binding sites per tannin molecule, the fraction of protein precipitated is:

where \([\text{T}]\) is tannin concentration and \(K_d\) is the dissociation constant. Typical values: \(K_d \approx 0.1\text{-}1\) mM,\(n \approx 2\text{-}4\).

Alkaloids: Neurotoxin Mimics

Alkaloids are nitrogen-containing secondary metabolites that typically interfere with animal neurotransmission by mimicking or blocking endogenous ligands:

• Nicotine (tobacco): agonist at nicotinic acetylcholine receptors (nAChR), causing persistent depolarization and eventual desensitization. Insect nAChR is 10-100x more sensitive than mammalian nAChR, explaining nicotine's efficacy as an insecticide.
• Caffeine (coffee, tea): competitive antagonist at adenosine A₂A receptors, blocking the sleep-promoting effect of adenosine.\(K_i = 12\) \(\mu\)M at A₂A. Also inhibits phosphodiesterase (PDE) at higher concentrations (\(K_i \approx 500\)\(\mu\)M), increasing cAMP levels.
• Strychnine (nux vomica): competitive antagonist at glycine receptors (GlyR) in the spinal cord, blocking inhibitory neurotransmission. The resulting unopposed excitation causes tetanic convulsions.\(K_i \approx 5\) nM at GlyR.
• Morphine (opium poppy): agonist at\(\mu\)-opioid receptors. Potent analgesic in mammals but may function in the plant as a nitrogen storage compound and antifungal agent.

Dose-Response: The Hill Equation

The pharmacological effect of an alkaloid toxin follows the Hill equation, which describes cooperative binding to a receptor:

where \(c\) is the toxin concentration, \(K_{1/2}\) (= EC₅₀) is the concentration producing 50% of maximum effect, and \(n\) is the Hill coefficient. For \(n = 1\), this reduces to simple Michaelis-Menten binding; \(n > 1\)indicates positive cooperativity (a steeper dose-response curve). The Hill coefficient reflects the sensitivity of the dose-response relationship:

Higher Hill coefficients mean a narrower concentration window between "no effect" and "full effect", creating a threshold-like response. This has important ecological implications: toxins with high cooperativity (\(n \geq 3\)) create sharp boundaries between "safe" and "lethal" doses.

Cyanogenic Glycosides: The "Cyanide Bomb"

Cyanogenic glycosides (linamarin in cassava, amygdalin in almonds, dhurrin in sorghum) are stable, non-toxic storage forms that release hydrogen cyanide (HCN) when the plant tissue is crushed. The "cyanide bomb" mechanism relies on compartmentalization:

HCN inhibits cytochrome c oxidase (Complex IV) in the mitochondrial electron transport chain with extreme potency:

The cyanide binds to the ferric iron in the binuclear center of Complex IV, preventing electron transfer to O₂ and halting aerobic respiration. The lethal dose in mammals is approximately 1-3 mg HCN/kg body weight (comparable to the amount released by 100 g of bitter cassava root).

2.3 Co-evolutionary Arms Races

Ehrlich & Raven (1964) proposed that the extraordinary diversity of plants and herbivorous insects is driven by reciprocal evolutionary escalation — what Van Valen (1973) later termed the "Red Queen" hypothesis: species must continuously evolve merely to maintain fitness relative to co-evolving antagonists.

Monarch-Milkweed: A Molecular Case Study

Milkweeds (Asclepias spp.) produce cardiac glycosides (cardenolides) — steroid derivatives that inhibit the Na⁺/K⁺-ATPase pump by binding to its extracellular surface. The most studied cardenolide, ouabain, binds with nanomolar affinity:

Inhibition of the Na⁺/K⁺-ATPase disrupts the electrochemical gradient across cell membranes, leading to cardiac arrhythmia and death in most animals. However, monarch butterflies (Danaus plexippus) have evolved a remarkable counter-adaptation: a single amino acid substitution in the \(\alpha\)-subunit of the Na⁺/K⁺-ATPase at position 122 (asparagine → histidine, N122H) reduces ouabain sensitivity by ~200-fold:

This mutation has evolved independently at least four times in cardenolide-feeding insects (monarchs, oleander aphids, large milkweed bugs, and a chrysomelid beetle), representing a stunning example of convergent molecular evolution under positive selection.

Mutation Fixation Under Positive Selection

The rate of fixation of an advantageous mutation depends on the population size and the selection coefficient. For a new mutation with selective advantage \(s\) in a diploid population of size \(N\), Kimura (1962) showed that the fixation probability is:

The expected time to fixation, conditional on fixation occurring, is approximately:

For the N122H mutation in monarchs, with estimated \(s \approx 0.01\) and\(N \approx 10^8\), this predicts \(\bar{t}_\text{fix} \approx 3800\)generations (~3,800 years given annual generation time). The mutation rate at any specific nucleotide is approximately \(\mu \approx 10^{-8}\) per generation, so the waiting time for the mutation to appear is \(1/(2N\mu) \approx 1/(2 \times 10^8 \times 10^{-8}) = 0.05\)generations — effectively instantaneous. The rate-limiting step is fixation, not mutation supply.

The Escalation Pattern

Ehrlich & Raven proposed that co-evolutionary arms races proceed through a series of escape-and-radiate cycles:

1. Plant innovation: a plant lineage evolves a novel chemical defense (e.g., cardenolides in Asclepiadaceae), escaping from most herbivores
2. Adaptive radiation of plants: freed from herbivore pressure, the defended lineage diversifies into empty ecological space
3. Herbivore counter-adaptation: an herbivore lineage evolves resistance (e.g., N122H mutation), gaining exclusive access to the defended plant lineage
4. Adaptive radiation of herbivores: the resistant herbivore lineage diversifies on the now-accessible plant radiation
5. The cycle repeats with a new plant defense innovation

Phylogenetic analyses confirm this pattern: major radiations in Lepidoptera (butterflies and moths) correlate with the evolution of detoxification mechanisms for specific plant chemical classes (Wheat et al. 2007). The Papilionidae (swallowtails) diversified after evolving CYP6B cytochrome P450 enzymes that detoxify furanocoumarins; the Pieridae (whites and sulphurs) radiated after evolving nitrile-specifier proteins that redirect glucosinolate hydrolysis away from toxic isothiocyanates.

2.4 Specialist vs Generalist Feeders

Herbivorous animals span a continuum from extreme specialists (monophages feeding on a single plant species) to broad generalists (polyphages feeding on dozens of plant families). This variation reflects a fundamental trade-off in detoxification capacity: investing in specific, high-affinity detoxification enzymes vs maintaining broad-spectrum detoxification at lower efficiency.

Specialist Detoxification: CYP450 Enzymes

Cytochrome P450 monooxygenases (CYP450s) are the primary enzymes for phase I detoxification of plant allelochemicals. These heme-containing enzymes catalyze oxidation reactions:

Specialist herbivores typically possess a small number of highly efficient, substrate-specific CYP450s. For example, Papilio polyxenes (black swallowtail) has CYP6B1 that metabolizes linear furanocoumarins with remarkable efficiency:

This catalytic efficiency approaches the diffusion limit for some substrates, indicating strong selective optimization.

Generalist Detoxification: Glutathione Conjugation

Generalist herbivores rely more heavily on broad-spectrum phase II detoxification via glutathione S-transferases (GSTs):

GSTs conjugate electrophilic toxins (epoxides, quinones, \(\alpha,\beta\)-unsaturated carbonyls) with glutathione (GSH), rendering them water-soluble for excretion. The advantage is broad substrate specificity; the cost is lower catalytic efficiency per substrate (\(k_\text{cat}/K_m \approx 10^2\text{-}10^4\) M⁻¹s⁻¹) and higher GSH consumption.

Optimal Diet Breadth: Marginal Value Theorem

The optimal number of plant species to include in the diet can be derived from the marginal value theorem (Charnov 1976). An herbivore searching for food encounters plant species \(i\) at rate\(\lambda_i\) and gains energy \(e_i\) with handling time\(h_i\). The average energy gain rate when feeding on the top \(k\)species (ranked by profitability \(e_i/h_i\)) is:

The optimal diet includes species \(k+1\) if and only if adding it increases the average rate:

That is, include a species if its profitability exceeds the current average. Crucially, the optimal diet breadth depends only on encounter rates and profitabilities — not on the abundance of any individual species. This explains why generalists often ignore low-quality food items even when they are abundant, and why specialists maintain narrow diets even when alternative foods are available (if the specialist detoxification advantage makes the primary host highly profitable).

The trade-off between specialist and generalist strategies also involves neural constraints: learning and recognizing many host plants requires larger neural investment (mushroom bodies in insects), creating a brain-size / diet-breadth trade-off documented in Lepidoptera (Bernays 2001).

Co-evolutionary Arms Race Spiral

The Ehrlich-Raven co-evolutionary model describes reciprocal adaptation cycles between plants and herbivores. Each escalation step involves novel chemistry — new defense compounds from plants, new detoxification enzymes from herbivores.

Python: Arms Races, Nectar, Toxins & Diet Breadth

This four-panel simulation explores: (1) co-evolutionary Red Queen dynamics showing plant defense and herbivore detoxification levels escalating over 500 generations with stochastic noise; (2) optimal nectar concentration from Poiseuille flow for different proboscis lengths, showing the viscosity-energy density trade-off; (3) dose-response curves for four plant toxins using the Hill equation, illustrating different potencies and cooperativities; and (4) optimal diet breadth from the marginal value theorem, showing how energy gain rate varies with the number of plant species in the diet.

2.5 Detoxification Biochemistry in Herbivores

Herbivores that feed on chemically defended plants must detoxify, sequester, or excrete the ingested allelochemicals. The detoxification machinery is organized into three sequential phases, each with distinct biochemical roles, collectively forming an integrated xenobiotic defense system.

Phase I: CYP450 Oxidation

Cytochrome P450 monooxygenases (CYP450s) catalyze the oxidative functionalization of lipophilic toxins, introducing a hydroxyl group that increases water solubility and provides a handle for Phase II conjugation:

The catalytic cycle involves the iron center of the heme prosthetic group cycling between Fe³⁺ and Fe²⁺ states, with molecular oxygen being activated and split — one oxygen atom is inserted into the substrate while the other is reduced to water. CYP450s are remarkably versatile, catalyzing hydroxylation, epoxidation, N-dealkylation, O-dealkylation, and deamination reactions.

Phase II: Conjugation Reactions

Phase II enzymes attach polar moieties to the Phase I products (or directly to toxins with suitable functional groups), dramatically increasing water solubility for excretion:

• Glucuronidation (UDP-glucuronosyltransferases): adds glucuronic acid — the most important conjugation pathway in mammals
• Sulfation (sulfotransferases): adds a sulfate group from PAPS (3'-phosphoadenosine-5'-phosphosulfate)
• Glutathione conjugation (glutathione S-transferases): adds glutathione (GSH) to electrophilic toxins — critical for detoxifying reactive intermediates from CYP450 oxidation
• Amino acid conjugation: adds glycine or other amino acids to carboxylic acid groups

Phase III: ABC Transporter Efflux

ATP-binding cassette (ABC) transporters actively pump conjugated metabolites out of cells using ATP hydrolysis. These membrane proteins recognize the Phase II conjugates (glucuronides, sulfates, glutathione conjugates) and transport them into the gut lumen, bile, or hemolymph for excretion. Key families include P-glycoprotein (MDR1/ABCB1) and the multidrug resistance-associated proteins (MRP/ABCC).

Clearance Rate and Detoxification Kinetics

The rate at which a herbivore can clear an ingested toxin follows Michaelis-Menten kinetics. The clearance rate CL describes the volume of body fluid cleared of toxin per unit time:

where \(V_\text{max}\) is the maximum detoxification rate, \(K_m\) is the Michaelis constant, and \([S]\) is the toxin concentration. At low toxin concentrations (\([S] \ll K_m\)), clearance is approximately constant:\(\text{CL} \approx V_\text{max}/K_m\) (first-order kinetics). At high concentrations (\([S] \gg K_m\)), the detoxification system saturates and clearance drops: \(\text{CL} \approx V_\text{max}/[S]\) (zero-order kinetics). This saturation is when toxicity becomes dangerous — the herbivore can no longer keep up with toxin intake.

The Arms Race Escalation in CYP450 Gene Families

The CYP450 superfamily provides one of the clearest molecular records of co-evolutionary arms race escalation. Herbivore CYP450 gene families have undergone repeated gene duplication and neofunctionalization in response to plant defense innovation. Key examples:

• CYP6B subfamily in Papilionidae: expanded from 2 to 12+ genes specifically for furanocoumarin detoxification — each paralog specialized for different furanocoumarin structures
• CYP6AE14 in Helicoverpa armigera(cotton bollworm): specifically detoxifies gossypol, the main defense compound in cotton
• CYP321A1 in Helicoverpa zea: broad-substrate CYP450 induced by many plant allelochemicals, representing a generalist strategy

Phylogenetic analysis reveals that CYP450 gene duplication events in herbivore lineages are temporally correlated with the evolution of novel defense compounds in their host plants, providing molecular evidence for the Ehrlich-Raven escape-and-radiate model.

2.6 Pollination Syndromes: Chemistry Matches Pollinator

Pollination syndromes are suites of floral traits — color, scent, nectar composition, flower shape, and timing of anthesis — that have converged across unrelated plant lineages in response to selection by specific pollinator guilds. The chemistry of each syndrome reflects the sensory biology, morphology, and energetics of the target pollinator.

Major Pollination Syndromes

Energetic Optimum: Nectar Concentration x Viscosity Trade-off

Each pollinator morphology has a different optimal nectar concentration, determined by the trade-off between energy density and viscosity. Recall from Poiseuille flow that energy intake rate is:

where \(L\) is the effective feeding apparatus length and \(\beta\)characterizes the viscosity-concentration relationship. The optimal concentration\(c^* = 1/\beta\) is independent of \(L\), but the absolute energy intake rate decreases with \(L\). This means that long-tongued pollinators (moths, hummingbirds) are more severely penalized by viscous nectar, which explains why:

• Bee-pollinated flowers produce the most concentrated nectar (30-50%) — short proboscis tolerates high viscosity
• Hummingbird flowers produce dilute nectar (20-25%) — long bill amplifies viscosity penalty
• Bat flowers produce the most dilute nectar (15-20%) — lapping mechanism is different from suction but still viscosity-limited

The "Honest Signal" Hypothesis

The honest signal hypothesis proposes that costly floral scent production serves as a reliable indicator of nectar reward quality. Scent compounds — particularly terpenoids and benzenoids — require significant metabolic investment (6+ ATP per molecule of linalool). Plants that invest heavily in scent production are signaling their capacity to also invest in nectar production. Pollinators that preferentially visit strongly scented flowers are, on average, rewarded with more nectar, reinforcing the correlation between signal intensity and reward quality.

This creates an evolutionary stable equilibrium: cheaters (plants producing strong scent but little nectar) are visited less frequently as pollinators learn to discriminate, maintaining the honesty of the signal. The cost of scent production thus functions as a Zahavian handicap — only plants with sufficient resources can afford both the signal and the reward.

2.7 Seed Dispersal Chemistry

The chemistry of fruits is shaped by the need to attract appropriate seed dispersers (typically birds and primates) while deterring seed predators (typically mammals that destroy seeds by mastication). This dual selective pressure has produced some of the most elegant chemical signaling systems in nature.

Fruit Color Signals Ripeness

The dramatic color change during fruit ripening serves as a visual signal to dispersers that seeds are mature and the fruit reward is available. The biochemical basis involves a coordinated pigment transition:

• Chlorophyll degradation: chlorophyllase and pheophytinase break down the green pigment, unmasking underlying carotenoids
• Carotenoid biosynthesis: \(\beta\)-carotene, lycopene, and xanthophylls are synthesized de novo, producing yellow, orange, and red colors (tomato: lycopene gives the characteristic red)
• Anthocyanin accumulation: flavonoid-derived pigments provide red, purple, and blue colors in many fruits (grapes, blueberries, cherries) through vacuolar pH-dependent color shifts

The specific color produced is matched to the visual system of the target disperser: bird-dispersed fruits are typically red or black (matching avian tetrachromatic vision), while primate-dispersed fruits are often yellow or orange (matching trichromatic vision shared by Old World primates and humans).

Capsaicin: Directed Deterrence

Capsaicin (8-methyl-N-vanillyl-6-nonenamide) in chili peppers (Capsicum spp.) provides a remarkable case study in targeted chemical defense. Capsaicin specifically deters mammals (which crush seeds with molars, destroying them) while having no effect on birds (which swallow seeds whole and disperse them intact).

The molecular basis lies in the TRPV1 receptor(transient receptor potential vanilloid 1), a heat- and pain-sensing ion channel. Capsaicin binds to the vanilloid binding pocket of mammalian TRPV1, causing the channel to open and triggering an intense burning sensation. The binding follows a standard dose-response relationship:

The key evolutionary innovation is that bird TRPV1 receptors are insensitive to capsaicin. In avian TRPV1, the vanilloid binding pocket has different amino acid residues (particularly at positions equivalent to mammalian S512 and T550), preventing capsaicin from binding. The EC₅₀ for capsaicin activation of bird TRPV1 is approximately 1,000-10,000 fold higher than for mammalian TRPV1:

This enormous difference in sensitivity means that the capsaicin concentrations found in ripe chili peppers (typically 10-100 \(\mu\)M in the pericarp) are more than sufficient to deter mammals (producing intense pain) while being far below the threshold for avian detection. The result is a chemical filter that selectively permits seed dispersal by birds while excluding seed predation by mammals — a textbook example of directed deterrence.

Dose-Response Across Taxa

The dose-response relationship for capsaicin varies not only between mammals and birds but also among mammalian species. Tree shrews (Tupaia belangeri), which are evolutionary intermediates between insectivores and primates in Southeast Asia, have a mutant TRPV1 with reduced capsaicin sensitivity, allowing them to consume wild chili peppers as a food source. This suggests that capsaicin sensitivity is an evolvable trait, and that some mammals have begun counter-adapting to this plant defense — another chapter in the ongoing chemical arms race.

Python: Detoxification, Pollination Syndromes & Capsaicin

This three-panel simulation explores: (1) CYP450 detoxification kinetics for multiple plant toxin substrates, showing Michaelis-Menten curves with different \(V_\text{max}\)and \(K_m\) values reflecting substrate specificity of specialist vs generalist detoxification enzymes; (2) pollination syndrome nectar optimization comparing the optimal sugar concentration for bees, butterflies, hummingbirds, and bats based on the Poiseuille flow viscosity-energy trade-off; and (3) capsaicin dose-response curves across mammalian and avian species, demonstrating the >1000-fold difference in TRPV1 sensitivity that enables directed seed dispersal.

References

1. Ehrlich, P.R. & Raven, P.H. (1964). Butterflies and plants: a study in coevolution. Evolution, 18(4), 586-608.
2. Van Valen, L. (1973). A new evolutionary law. Evolutionary Theory, 1, 1-30.
3. Kingsolver, J.G. & Daniel, T.L. (1983). Mechanical determinants of nectar feeding strategy in hummingbirds: energetics, tongue morphology, and licking behavior. Oecologia, 60(2), 214-226.
4. Dobler, S., Dalla, S., Wagschal, V. & Agrawal, A.A. (2012). Community-wide convergent evolution in insect adaptation to toxic cardenolides by substitutions in the Na,K-ATPase. Proceedings of the National Academy of Sciences, 109(32), 13040-13045.
5. Charnov, E.L. (1976). Optimal foraging, the marginal value theorem. Theoretical Population Biology, 9(2), 129-136.
6. Wheat, C.W., Vogel, H., Wittstock, U., Braby, M.F., Underwood, D. & Mitchell-Olds, T. (2007). The genetic basis of a plant-insect coevolutionary key innovation. Proceedings of the National Academy of Sciences, 104(51), 20427-20431.
7. Bernays, E.A. (2001). Neural limitations in phytophagous insects: implications for diet breadth and evolution of host affiliation. Annual Review of Entomology, 46(1), 703-727.
8. Kimura, M. (1962). On the probability of fixation of mutant genes in a population. Genetics, 47(6), 713-719.
9. Hill, A.V. (1910). The possible effects of the aggregation of the molecules of haemoglobin on its dissociation curves. Journal of Physiology, 40, iv-vii.
10. Agrawal, A.A. & Fishbein, M. (2006). Plant defense syndromes. Ecology, 87(sp7), S132-S149.
11. Li, X., Schuler, M.A. & Berenbaum, M.R. (2007). Molecular mechanisms of metabolic resistance to synthetic and natural xenobiotics. Annual Review of Entomology, 52, 231-253.
12. Fenster, C.B., Armbruster, W.S., Wilson, P., Dudash, M.R. & Thomson, J.D. (2004). Pollination syndromes and floral specialization. Annual Review of Ecology, Evolution, and Systematics, 35, 375-403.
13. Tewksbury, J.J. & Nabhan, G.P. (2001). Directed deterrence by capsaicin in chillies. Nature, 412(6845), 403-404.
14. Caterina, M.J., Schumacher, M.A., Tominaga, M., Rosen, T.A., Levine, J.D. & Julius, D. (1997). The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature, 389(6653), 816-824.
15. Jordt, S.E. & Julius, D. (2002). Molecular basis for species-specific sensitivity to "hot" chili peppers. Cell, 108(3), 421-430.
16. Willmer, P. (2011). Pollination and Floral Ecology. Princeton University Press.